structural transformation options of the saudi economy
TRANSCRIPT
Structural Transformation Options of the Saudi Economy Under
Constraint of Depressed World Oil Prices
Salaheddine Soummanea, FrΓ©dΓ©ric Ghersib, Franck Lecocqc
a Corresponding author. CIRED, AgroParisTech, CIRAD, CNRS, EHESS, Ecole des Ponts ParisTech, Univ Paris-Saclay, UMR 8568. CIRED,
94736 Nogent-sur-Marne, France. E-mail: [email protected].
b CIRED, AgroParisTech, CIRAD, CNRS, EHESS, Ecole des Ponts ParisTech, Univ Paris-Saclay, UMR 8568. CIRED, 94736 Nogent-sur-
Marne, France. E-mail: [email protected].
c CIRED, AgroParisTech, CIRAD, CNRS, EHESS, Ecole des Ponts ParisTech, Univ Paris-Saclay, UMR 8568. CIRED, 94736 Nogent-sur-
Marne, France. E-mail: [email protected].
ABSTRACT
We implement the hybrid (energy-economy) recursive-dynamic multisector IMACLIM model with important
adaptations to Saudi macroeconomics. We design two scenarios reflecting both the Saudi Vision 2030 economic
development program and Nationally Determined Contribution (NDC) to greenhouse gas mitigation: Continuity
of previous plans to expand energy-intensive activities under maintained energy-pricing policies, versus
Transformation by economic diversification away from hydrocarbon-related activities and fiscal and energy-
pricing reforms. We show that, compared to Continuity, Transformation improves activity, employment and public
budget outlooks, while considerably abating the energy intensity of GDP and total CO2 emissions. Our results thus
point at the relevance of economic diversification as both a hedging strategy against international climate change
mitigation depressing oil markets and a national climate mitigation strategy for Saudi Arabia. However, the
successful advancement of the reforms necessary for diversification remains conditional to setting a suitable
institutional framework for a competitive economy.
KEYWORDS
Economic diversification, General equilibrium, Oil-exporting country, Climate policy.
2
1 INTRODUCTION
The implications of global mitigation policies for oil-exporting countries of the Middle East have generated
considerable attention. There is a large consensus in the literature that a scenario of ambitious mitigation would
depress oil prices. The drop is of 27% by 2030 in the Sustainable Development Scenario (SDS) of the International
Energy Agency (IEA, 2017) compared to its less stringent New Policies Scenario (NPS). There is also widespread
agreement that mitigation policies at global scale will reduce global oil demand. Comparatively, how demand for
the conventional oil of the Middle East and thus how export revenues of Middle East producers may evolve, is
less clear-cut. Johansson et al. (2009) and Persson et al. (2007) project gains as carbon pricing would drive
unconventional resources out of the market under the condition of low oil-demand elasticity. Conversely, Bauer
et al. (2016), Edenhofer et al. (2014), Waisman et al. (2013) estimate that oil-exporting countries in the Middle
East will lose from the implementation of global mitigation policies, as will all oil-exporters.
The Kingdom of Saudi Arabia (KSA) is notoriously exposed to both volume and valuation risks, as oil revenues
generated over 80% of both the countryβs exports and fiscal earnings between 2009 and 2018 (SAMA, 2018).
Consequently, the KSAβs communications to the United Nations Framework Convention on Climate Change
(UNFCCC) have argued that global climate policies would deteriorate its economic outlook (Saudi Arabia, 2005;
2011; 2016a). Some strategies could mitigate adverse climate-policy effects on oil revenues (e.g., quota
agreements to sustain prices, price cuts to increase market shares or monetary compensation under the climate-
negotiation process) but may prove challenging to implement and transitory at best. Pershing (2000), Hvidt (2013)
and Van de Graaf and Verbruggen (2015) affirm that economic diversification is the only viable development
strategy for oil-exporting countries facing the structural decline of oil rent prompted by global low-carbon
transition.
Economic diversification away from oil extraction has been a constant objective of KSA policymakers since the
1970s (Saudi Arabia, 2016a). Most plans have focused on the development of energy-intensive industries to exploit
the countryβs competitive advantage on energy costs. The latest planβcalled Vision 2030βoutlines a more
comprehensive transformational strategy to drive the Saudi economy away from oil dependence (Saudi Arabia,
2016b). Although Vision 2030 expects oil and gas extraction to maintain an important role in the Saudi economy,
the transformative aspect of the plan sets the relative weight of that activity to decrease as investment flows towards
non-oil sectors. Besides, the plan orients diversification less towards energy-intensive industries and more towards
3
services. Additionally, diversification beyond energy-intensive activities is the cornerstone of the KSA Nationally
Determined Contribution (NDC) to the Paris Agreement (Saudi Arabia, 2015). However, Saudi NDC also
considers an alternative scenario of limited diversification focused on energy-intensive industries.
The objective of the present paper is to assess the economic, fiscal, social and environmental implications of
diversification of the Saudi economy under two scenarios reflecting those of the Saudi NDC. The test of two
strategies is justified by the fact that successive development plans have so far failed to foster the emergence of a
competitive non-oil private sector (Albassam, 2015), and that the latter objective continues facing considerable
challenges, mostly regarding job creation (IMF, 2016a).
The methodological challenge of modeling the Saudi economy has been taken up in recent years. Al-Thumairi
(2012), Al-Hawwas (2010) and De Santis (2003) assess static counterfactual scenarios, mostly of oil price shocks,
with conventional computable general equilibrium (CGE) models. Blazquez et al. (2017) implement a stylized
dynamic equilibrium model to estimate the long-run welfare impact of renewable energy penetration. Gonand et
al. (2019) model overlapping generations in a similar dynamic framework to explore the long-run consequences
of energy pricing reforms.
Compared to this corpus, our contribution is to implement a dynamic economy-wide model adapted to KSA
macroeconomics, calibrated both statically and dynamically on hybrid economy/energy data and backed by
bottom-up energy modeling (Ghersi, 2015; Hourcade et al. 2006). In this spirit, Soummane et al. (2019) developed
the KLEM-KSA model to investigate the macroeconomics of reduced Saudi oil-export income at the 2030 horizon.
They showed that structural weakening of the oil price would slow down Saudi growth and increase
unemployment, while significantly cutting down trade-surplus accumulation and increasing the public debt. They
then demonstrated how domestic energy-pricing reforms could mitigate some of these losses if they were partly
recycled in higher investment, irrespective of their impact on energy intensity. The level of aggregation of their
analysis only allowed for crude descriptions of both production activities (two aggregate sectors) and the secondary
distribution of income (rough estimates of public budget balances, no assessment of debt accumulation).
Comparingly, our modeling breaks down energy supply into four sectors and non-energy activity into nine sectors,
thereby allowing exploration of diversification options. It also extends to the secondary distribution of income
between households, firms, public administrations and the rest of the world, thus producing original outlooks on
the Saudi national debt and its distribution across domestic agents. With this tool, we produce dynamic outlooks
4
at medium-term horizons of yet unexplored macroeconomic scenarios combining sectoral diversification, energy-
pricing and public budget reforms.
The rest of our paper is organized as follows. We present our modeling method in Section 2. In Section 3, we
describe two mitigation scenarios inspired by the Saudi NDC, then comment on scenario results in Section 4. We
conclude and highlight policy implications in Section 5.
2 THE IMACLIM-SAU MODEL
2.1 Salient features
IMACLIM is an economy-wide model representing the supply and demand of goods and services, with the specific
purpose of articulating with engineering representations of the energy system to produce consistent energy-
economy outlooks (Ghersi, 2015; Hourcade et al., 2006). IMACLIM exists in a global multi-regional version
(Crassous et al., 2006; Sassi et al., 2010) and in country-specific versions currently covering France (Hourcade et
al., 2010; Le Treut, 2017; De Lauretis, 2017), South Africa (Schers et al., 2015), Brazil (Lefèvre et al., 2018) and
India (Gupta et al., 2019, 2020).
The application of IMACLIM to Saudi Arabia, IMACLIM-SAU, is a dynamic simulation model picturing
economic growth in yearly time steps as resulting from exogenous assumptions of labor endowment and
productivity (the Harrod-neutral assumption on technical progress). The vector of domestic outputs at year π‘, ππ‘ ,
is a function of the stock of capital πΎπ‘, of the labor force πΏπ‘ , and of the intermediate consumption of energy and
non-energy resources. Capital stock dynamics follow the standard accumulation rule πΎπ‘+1 = (1 β πΏ) πΎπ‘ + πΌπ‘, with
πΏ the depreciation rate, constant over the modeled horizon. Investment πΌπ‘ is the amount of non-energy output used
to build up πΎ at period π‘. Beyond these core Solow-Swann specifications and similar to other country versions,
IMACLIM-SAU deviates from the CGE standard by four salient features.1
Firstly, IMACLIM-SAU calibrates on original hybrid data reconciling national accounting statistics with energy
flows and prices statistics. This allows maintaining dual accounting of energy flows in physical and monetary units
1 Annex A in the Online Appendix presents the formulary and reference tables of the model. Section 2, Annex A and Annex B in the Online
Appendix draw from Soummane and Ghersi (2020), with a few corrections and modifications.
5
backed by statistically relevant, agent-specific prices. Hybrid accounting has a significant bearing on benchmark
ratios crucial to macroeconomic analysis: the cost shares of energy in productions, the budget share of energy for
households and the breakdown of energy consumptions and CO2 emissions across sectors and agents (Combet et
al., 2014; Le Treut, 2017).
Secondly, IMACLIM-SAU treats as exogenous all variables pertaining to the energy system. This feature stems
from the model being designed to couple with bottom-up energy expertise (Ghersi, 2015). IMACLIM-SAU thus
traces growth trajectories building around exogenous energy flows. The cost structure of energy supply beyond its
own energy intensity, as well as the specific margins on all energy sales, are also exogenously adjusted to match
assumptions on the dynamics of annualized investment, operational expenses, domestic consumer prices and trade
prices. Such specifications allocate part of value-added to constrained energy expenses and part of primary factor
endowments to constrained energy supply volumes. These constraints on volumes, costs, and prices weigh on
economic growth.
Thirdly, IMACLIM-SAU simulates suboptimal growth pathways. One first deviation from optimal growth is that
IMACLIM-SAU is a simulation model that builds on exogenous investment pathways not explicitly reflecting the
intertemporal optimization of some welfare indicator. A second deviation is that the model, on top of exogenously
constraining energy markets, considers imperfect non-energy markets in the form of both mark-up pricing and
underutilization of labor. On the labor market, the inertia of real wages prevents full clearing i.e. induces
equilibrium unemployment specified through a βwage curveβ correlating the unemployment rate and the real wage
(Blanchflower and Oswald, 2005). The specification captures well the characteristic rigidity of Saudi wages
(Devaux, 2013).
Lastly, IMACLIM-SAU is carefully adapted to Saudi macroeconomics. These are uncommonly specific regarding
both investment and saving behavior, and the trade balance. On investment and savings, the use of the sovereign
Saudi fund as a buffer against fluctuations of oil export revenues (AlKathiri et al., 2020) requires departing from
the CGE standard of fixed savings driving investment. Variations of the sovereign fund amount to variations of
national savings meant to smooth out the investment path of the Kingdom. IMACLIM-SAU reflects this
macroeconomic policy by implementing the Johansen closure of national savings adjusting to some exogenous
investment path, rather than the neoclassical closure of investment adjusting to national savings (Taylor and Lysy,
1979).
6
On the trade balance, both the currency peg that has been fixing the value of the Saudi Riyal (SAR) against that of
the US dollar since 1986 and the sensitivity of Saudi exports to the global oil price forbid considering the
neoclassical standard of an exogenous trade balance resulting from real effective exchange rate (REER)
adjustments. One alternate specification is to consider a fixed ratio between Saudi and import prices as a proxy of
the fixed exchange rate to the USD. However, statistics reveal that the REER of Saudi Arabia is not fixed, but
correlated to the contribution of the trade balance to Saudi GDP. The reason for this correlation is the differentiated
impacts of oil price fluctuations on Saudi versus foreign (especially the US) prices (Soummane et al., 2019).
IMACLIM-SAU substitutes the correlation to the standard trade-balance constraint for those scenarios considering
the maintained administration of domestic energy prices, i.e. our Continuity scenario (see Section 3 and Equation
A-32a). Scenarios of energy-pricing reforms like our Transformation scenario, because they modify the Saudi
REER through unilateral increases of production costs, require yet another trade specification. We follow the
reasoning of Soummane et al. (2019)βs Annex B by dropping the correlation between the REER and the trade
balance and forcing the ratio of the βrental priceβ of domestic value-added to foreign prices of our Continuity
scenario, in our Transformation scenario (Equation A-32b).
2.2 Model calibration
We calibrate IMACLIM-SAU at its 2013 base year on the original βhybridβ energy-economy dataset of Soummane
and Ghersi (2019). Additionally, to improve the relevance of modeled trajectories, we perform dynamic calibration
for the years 2014 to 2017 targeting GDP, unemployment, and trade balance statistics, under the constraint of
observed investment efforts, energy flows, and prices.
The latter calibration targets macroeconomic variables common to IMACLIM-SAU and the compact KLEM-KSA
model of Soummane et al. (2019). We therefore follow Soummane et al.βs (2019) procedure and refer to their
Annex D.2 for its description. The static calibration is similar to that of KLEM-KSA but extends in two directions:
the breakdown of activity in 13 sectors, four of which are energy sectors (Table 1), and the secondary distribution
of income between households, firms, public administrations and the foreign βrest-of-the-worldβ agent (see Annex
B in the Online Appendix).
7
Table 1: IMACLIM-SAU sectors
Abbreviation Goods and services Specification
OIL Crude oil Hybrid
GAS Natural gas Hybrid
REF Refined petroleum products Hybrid
ELE Electricity Hybrid
AGR Agriculture, hunting, forestry, and fishing products Non-hybrid
MIN Other mining products Non-hybrid
CHM Chemicals and petrochemicals Non-hybrid
NMM Non-metallic minerals (including cement) Non-hybrid
MAN Manufactured products Non-hybrid
C&S Construction and services (commercial and public) Non-hybrid
WTP Water transport services Non-hybrid
ATP Air transport services Non-hybrid
OTP Other transport services Non-hybrid
βHybridβ sectors rest on calibration data extending to satellite accounts of physical flows, which imply agent-specific prices (see Section
2.1 and Soummane et al., 2019).
3 SCENARIOS NARRATIVES AND PARAMETRIZATION
Considering its level of sectoral and distributional detail, we limit our exploration to the 2030 horizon, where we
can both sustain its parameterization with a mix of governmental and expert views, and assume prolongation of
KSA macroeconomics, particularly the currency peg to the USD. At that horizon, we focus on two contrasted
scenarios.
3.1 Energy-intensive development with limited diversification: The Continuity scenario
Since the establishment of the Royal Commission for Jubail and Yanbu (RCJY) in 1975, the KSA has invested
heavily in energy-intensive industries. The expansion of such industries benefited from two comparative
advantages, namely integration with oil and gas extraction and low regulated energy prices.2 In 2004, 2009 and
2015, the RCJY successively grew with the Jubail II and Yanbu II projects, the Ras Al Khair city dedicated to
2 For example, Moya and Boulamanti (2016) report that the price of natural gas for some industrial activities in Saudi Arabia is one tenth what
it is in Europe.
8
mining activities and the Jazan city dedicated to downstream activities. The Vision 2030 program, although a
strategic plan to reduce the Kingdomβs reliance on oil, still identifies activities related to hydrocarbons as pillars
of the economy producing revenue crucial to the planβs success.
Our Continuity scenario prolongs this trend of strong development of energy-intensive industries under national
climate policy action of lesser ambition, corresponding to the alternative scenario (βscenario 2β) of the Saudi NDC
(Saudi Arabia, 2015). It assumes unchanged energy-price regulations resulting in weak energy-efficiency gains
(see Annex C in the Online Appendix) combined to export stimuli as main policy drivers of the further expansion
of energy-intensive industries. We translate the latter assumption in IMACLIM-SAU by acting on exogenous
export trends (see Equation A-30 of Annex A in the Online Appendix), which modify sectoral activity and
consequently, investment flows. We focus export-trend boosts on three sectors explicitly mentioned in βscenario
2β of the Saudi NDC, namely petrochemicals (CHM), mining (MIN) and cement (NMM).3
Saudi petrochemical exports grew rapidly in recent years and currently account for around 60% of non-oil exports
earnings (SAMA, 2018). Although their growth slowed down following the oil price slump of 2015 and 2016, it
recovered in 2017, recording a rate of +12.8%. This rate is only slightly above its +10.7 % annual average since
2005. Our Continuity scenario assumes that the latter average prevails up to 2030. The strong dynamics are
expected to proceed from the combination of increased current productions and the development of new
productions (Jadwa, 2017).
The MIN sector has significant untapped potential that the government is willing to mobilize (MGI, 2015). Recent
figures show spectacular increases of minerals outputs and exports.4 Our Continuity scenario assumes that MIN
exports grow at 7.1% per year between 2013 and 2030, reflecting the upper bound of MGI (2015).
Finally, the cement sector (NMM) has long been subject to an export ban, but the recent ban lift and the dynamics
of adjacent markets, especially those in current or foreseeable reconstruction phases (e.g., Iraq, Yemen, and Syria)
could boost the sectorβs exports. Our Continuity scenario therefore assumes NMM exports increasing at the same
7.1% annual rate as mineral exports.
3 Due to lack of data, we drop metal production as a distinct sector although it is cited in the Kingdomβs NDC.
4 https://www.export.gov/apex/article2?id=Saudi-Arabia-Mining-and-Minerals.
9
3.2 Proactive economic diversification compatible with higher climate pledges: The
Transformation scenario
Our Transformation scenario builds on the success of horizontal diversification plans and their compatibility with
stronger climate commitments, i.e., reflects the central scenario of the Saudi NDC (Saudi Arabia, 2015) and the
Vision 2030 program (Saudi Arabia, 2016b). It relies on the reform of energy prices to reduce opportunity costs
incurred from domestic rather than international energy sales and to stimulate energy-efficiency gains (see Annex
C.1 in the Online Appendix), and on the promotion of manufacturing and service activities as main drivers of
transformation of the Saudi economy.
For decades, Saudi manufacturing has been facing competitiveness hurdles like technology gaps, inadequate labor
skills and the lack of dedicated infrastructures (El Mallakh, 2015). To overcome such barriers, the Saudi
government founded the Industrial Clusters (IC) state agency, which means to identify, support, and activate
industrial investment opportunities. National and regional contexts provide promising opportunities for the
automotive industry, which would integrate with related upstream branches such as high-grade plastics. HSBC
(2018) projects a dynamic annual increase for Saudi transport equipment, machinery and other manufactured
exports at around 10% per year on average from 2017 to 2030. We retain this accelerated trend for the exports of
the MAN sector of our Transformation scenario.
Turning to services, we consider tourism and finance given the Kingdomβs background and regional experiences.
Exports of touristic activities account for most of services exports and almost 6% of total exports in 2013 (CDSI,
2014). In the next decade, tourism is expected to grow rapidly, increasing its contribution to GDP, employment,
and trade (Al Yousif and Al Bakr, 2017). Indeed, the KSA plans to reach 30 million visitors for Umrah alone by
2030, up from the current eight million (Saudi Arabia, 2016b). Exports of insurance and financial services account
for only 17% of total Saudi services exports, compared with almost 50% for Bahrain,5 where they contributed to
the decreased dependence on hydrocarbon revenues (Callen et al., 2014). In our Transformation outlook, we
consequently set Construction and Services (C&S) exports to grow at an annual rate of 6.7%. This figure is close
5 Source: The World Bank statistics at https://data.worldbank.org/indicator/BX.GSR.INSF.ZS?view=chart.
10
to the average growth of religious tourism between 2005 and 2014 (MGI, 2015) and reflects its targeted increase
up to 2030 (Saudi Arabia, 2016b).
One additional dimension of Transformation regards public finances. The Saudi government currently derives
income mostly from companies of the energy and energy-intensive sectors (see Annex B.1 in the Online
Appendix). The Transformation scenario offers the opportunity to gradually disengage the government from direct
economic activity by empowering the private sector (Saudi Arabia, 2016a). Fiscal reforms must compensate for
this shift, which is an important dimension of the Kingdomβs economic modernization.6 Our Transformation
outlook consequently considers the gradual increase of corporate taxes from the calibration rate of 2.5% (the Zakat)
to the current world average of 23.0% (Tax Foundation, 2018) by 2030. Although the increase is substantial, the
2030 rate remains below the current 27.4% average across G20 countries.
3.3 Scenario parameterization
In IMACLIM-SAU, economic growth primarily proceeds from exogenous increases of labor endowment and
productivity. For these two drivers, both our scenarios replicate the assumptions of Soummane et al. (2019)
reflecting projections by the International Labor Organization and Oxford Economics. The resulting trend of
efficient-labor growth is +2.2% per year, compounding into a potential 45.7% increase of real GDP between 2013
and 2030 (Table 2).
We exposed in sections 3.1 and 3.2 how our scenarios rest on differentiated export trends. In each scenario, we
apply to all not-singled-out non-energy sectors a default trend equal to the aggregate export-trend assumption of
Soummane et al. (2019), which reflects growth projections for the main export markets of the Middle East and
North Africa by the IMF (2016b). The favorable trade impulses of target sectors must not be considered
incremental sources of growth. They are rather forced alternative allocations to activities of the countryβs primary
factor endowments, while the main growth driver of efficient-labor increases is unaffected by scenarios. The
compared merits of the two allocations hang on the compared input structures and ex-ante international
competitiveness of targeted sectors, which determine their abilities to concretize the potential growth embedded
in efficient-labor dynamics.
6 MGI (2015, p. 112) argues that the KSA should consider the βintroduction of a modern system of taxation, along the lines of fiscal systems
common to all G20 emerging economiesβ.
11
On energy, both scenarios share a common international oil-price trajectory consistent with ambitious global
reductions of CO2 emissions highlighting the challenge facing oil-dependent economies. Beyond dynamic
calibration years, i.e., from 2018 to 2030, we use the price trajectory of the Sustainable Development Scenario of
the International Energy Agency (IEA, 2017). By contrast, as presented in sections 3.1 and 3.2 scenarios diverge
on domestic energy prices. The Continuity scenario assumes maintained regulation at current real levels while the
Transformation scenario considers ambitious reforms (see Annex C.1 in the Online Appendix). They consequently
diverge on induced energy-efficiency gains (see Annex C.2 in the Online Appendix). Another distinction between
scenarios is the Saudi strategy concerning crude oil production and exports. In the Continuity scenario, we assume
that the KSA sticks to its projected output according to IEA (2017). Under this assumption, output reaches 12.7
million barrels per day by 2030, which is close to the current production capacity of the Kingdom. Oil exports
endogenously result from the difference between output and domestic consumption. In the Transformation
scenario, we choose to maintain oil exports at their levels of the Continuity scenario by adjusting output, taking
account of domestic demand. Thus, the KSA avoids any βmarket-floodingβ strategy (see Annex C.4 in the Online
Appendix) that might have negative effects on oil prices (Blazquez et al., 2017).
Lastly, both scenarios reflect the introduction by the Kingdom of a value-added tax (VAT) at 5% in 2018, which
they maintain up to 2030. Transformation additionally increases the corporate tax rate exponentially from its 2.5%
value of 2017 to 23% in 2030 (see Section 3.2). To account for potential feedbacks from contrasted public-budget
performances, we cap public surplus at 1% of GDP and assume that extra net public income is redirected to
supplemental investment.
12
Table 2: Main scenario parameters
Calibration ------Continuity------ ----Transformation----
Macroeconomic drivers 2013 2030 AAGRa 2030 AAGR
Labor endowment (index)b 1.00 1.38 +1.9% 1.38 +1.9%
Labor productivity 1.00 1.05 +0.3% 1.05 +0.3%
Efficient Labor (index)c 1.00 1.46 +2.2% 1.46 +2.2%
Structural change drivers 2013 2030 AAGR 2030 AAGR
Default export trendd 1.00 1.76 +3.4% 1.76 +3.4%
MIN export trend 1.00 3.20 +7.1% 1.76 +3.4%
CHM export trend 1.00 4.19 +8.8% 1.76 +3.4%
NMM export trend 1.00 3.20 +7.1% 1.76 +3.4%
MAN export trend 1.00 1.76 +3.4% 3.86 +8.3%
C&S export trend 1.00 1.76 +3.4% 3.00 +6.7%
Energy system parameters 2013 2030 AAGR 2030 AAGR
World oil price (2016 USD per barrel)
108 69 -2.6% 69 -2.6%
Domestic energy prices Regulated Unreformede Reformedf
Crude oil output (Mtoe) 454.4 625.6 Endogenous
Crude oil exports (Mtoe) 381.9 Endogenous Equal to Continuity
Energy efficiency - Low gainse High gainse
Fiscal parameters 2013 2030 AAGR 2030 AAGR
Corporate income tax rate 2.5% 2.5% 0% 23.0% 12.3%
Value-added tax rate 0% 5% (from 2018 on) 5% (from 2018 on)
a AAGR is the average annual growth rate from 2013 to 2030. Drivers other than structural change drivers have non-constant annual growth
rates available from the authors upon request.
b Labor endowment is reported as index following common CGE practice (see e.g., Rutherford, 1999) and to clarify the computation of the
efficient labor index (see below). It figures in full-time equivalents in the model (see Table A1 of online appendix).
c Efficient labor gains are the product of labor endowment and labor productivity gains.
d Default trend applies to non-energy sectors unless otherwise specified.
e See Annex C.1 in the Online Appendix.
f See Annex C.2 in the Online Appendix.
4 SCENARIO RESULTS
We implement the two scenarios of Section 3 in IMACLIM-SAU running in annual steps from 2013 to 2030. In
the following subsections, we report on macroeconomic indicators public accounts dynamics, energy
consumptions and carbon emissions. For reference purposes, we synthesize below the main simulation results at
end-horizon (Table 3). Annex D in the Online Appendix provides some sensitivity analysis of the Transformation
scenario.
13
Table 3: Summary of modeling results at 2030 horizon
Indicator Calibration
2013
Continuity
2030
Transformatio
n2030
2030 variation
from Cont. to
Transf.
Real GDP, Bn 2013 SAR 2,773 3,973 4,025 +1.3%
Trade balance, % GDP +24.6% +9.8% +5.6% -4.3 pts
Cumulated 2013-2030 trade surplus, Bn 2013 USD
- 1,183 994 -16.0%
Unemployment rate 5.6% 8.6% 6.9% -1.7 pts
Public budget balance,
% GDP +5.6% -5.2% +1.0% +6.2 pts
Net public debt, % GDP
-95.9% +14.9% -21.0% -35.9 pts
CO2 emissions, Mt 502 913 405 -55.6%
Sources: Calibration data (see Section 2.2), IMACLIM-SAU simulations. βPtsβ stands for percentage points.
4.1 Macroeconomics
The Continuity and Transformation scenarios project real GDP growth from 2013 to 2030 at 43.2% and 45.1%
respectively, i.e. 2.5 points below potential (efficient-labor) growth for the former scenario versus 0.6 points below
potential growth for the latter. The key to the performance of Transformation is in the changed sectoral allocation
of primary factors of production via structural change, as well as in the additional investment effort that its public
budget performance allows funding.
Transformation channels an increasing share of investment towards non-energy-intensive sectors. In 2030, capital
stocks of the MAN and C&S activities end up 12.6% and 6.2% above their levels under Continuity, respectively.
At that year, MAN and C&S mobilize 8.0% and 85.2% of the non-energy capital stock.7 The shares are respectively
0.7 percentage points (pts) and 3.5 pts lower under Continuity. Conversely, Continuity concentrates a higher share
of capital stock in CHM, at 7.5% of the capital of non-energy sectors in 2030 versus 3.4% under Transformation.
These investment dynamics reflect the forced evolution of aggregate demand via export boosts, which reorients
activity towards target sectors. Continuity increases the activity shares of MIN, CHM and NMM to 5.8% of GDP
7 The share of the capital stock mobilized in energy sectors stems from the exogenous energy trajectories backing each scenario (see Annex
C).
14
in 2030 against 3.1% under Transformation. By contrast, Transformation induces a higher contribution from MAN
and C&S, who increase the sum of their shares to 57.4% of Saudi GDP in 2030 versus 52.1% under Continuity.
Activity dynamics drive the scenariosβ consequences on employment. The MAN and C&S sectors targeted by
Transformation are labor-intensive compared to the CHM, NMM and MIN sectors targeted by Continuity.
Compared to Continuity, the surges of activity in the C&S and MAN sectors of the Transformation scenario
therefore create significantly more employment (+424.3 thousand jobs) than the decreases of CHM, NMM and
MIN activities destroy (-142.1 thousand jobs, see Table 4). The net difference totals 269.6 thousand jobs in 2030,
which translates into a 1.7-point gap between the unemployment rates of Transformation, at 6.9%, and Continuity,
at 8.6%.
Table 4: Net employment creation by sector, thousand full-time equivalents
Sector Continuity
2030 versus 2013
Transformation
2030 versus 2013
2030 variation from
Cont. to Transf.
OIL 28.6 16.3 -12.2
GAS 17.4 -0.5 -17.9
REF 11.2 10.4 -0.8
ELE 31.8 12.1 -19.7
AGR 101.1 142.3 41.3
MIN 2.8 0.8 -2.0
CHM 145.6 20.7 -124.9
NMM 77.6 62.3 -15.2
MAN 121.8 216.9 95.1
C&S 3107.7 3436.8 329.2
WTP 8.6 9.1 0.5
ATP 16.0 20.0 4.0
OTP 61.0 53.3 -7.7
Total 3,731.1 4,000.6 269.6
Source: IMACLIM-SAU simulations. Sector codes are those of Table 1. Numbers may not add up due to rounding.
General equilibrium effects must not be overlooked when interpreting such results. The higher labor intensities of
MAN and C&S imply lower average labor productivity under Transformation than under Continuity. Higher
employment cannot therefore be the cause, but can only be the consequence of higher activity. The cause lies in
the other primary-factor market, that of capital. Under Continuity, the high capital intensities of target sectors put
strong pressure on capital demand, which induces sharp increases of capital costs. Transformation orients growth
towards less capital-intensive sectors and secures enough public resources to allow increasing the investment effort
in later projection years (see Section 4.2 below). At end-horizon, the ratio of capital to labor costs is thus 11.5%
15
higher under Continuity than under Transformation, in part given higher real wages under Transformation. The
increase of capital costs hampers both the purchasing power of households and the competitiveness of Saudi
exports on international markets via producer prices. The adverse impact on activity is reinforced by the possibility
of equilibrium unemployment.
Figure 1: Trade balance contribution to GDP under Continuity versus Transformation
Sources: 2013 calibration value from SAMA (2018), 2013 to 2017 dynamics from authorsβ computations on International Monetary Fund
and The World Bank data. IMACLIM-SAU simulations from 2018 to 2030.
Concerning international trade, both scenarios gradually improve the trade balance (Figure 1) from its 2015 lowest
of -8.7% of GDP up to 2025 thanks to the oil-price recovery projected by the SDS scenario of the IEA, then degrade
it up to 2030 following the oil price decline (see Annex C.3 in the Online Appendix). The trade surplus ends up at
9.8% of Continuity GDP thanks to that scenarioβs impulse on energy-intensive exports. Exports of CHM increase
by 275.7% in real terms from 2013 to 2030 and more than triple their contribution to total exports at current prices
(28.8% in 2030 versus 8.9% in 2013). MIN and NMM exports increase by 193.6% and 192.6%. The gap to the
forced trends of +318.7% for CHM and +220.0% for MIN and NMM (see Table 2) betrays the negative impact of
the oil-price contraction below 2013 levels from 2015 on, on the competitiveness of Saudi products, which our
original trade specification duly captures (see Section 2.1).
Under Transformation, the trade surplus contribution to GDP is lower in later years and ends at 5.6% in 2030. The
gradually weakening dynamics reflect exports of energy-intensive sectors suffering from energy-pricing reforms
16
(see Annex C.1 in the Online Appendix). In 2030, the ratios of domestic to foreign prices of CHM and NMM
under Transformation are thus 10.8% and 17.7% above their levels under Continuity. Consequently, real exports
of MIN, CHM, and NMM grow by 54.1%, 44.5% and 39.7% from 2013 to 2030, respectively, i.e. significantly
less than their assumed trends of 76.6% outside relative-price shifts (see Table 2). Yet, the trade balance of the
Transformation scenario benefits from a higher contribution of refined products, resulting from domestic energy
savings redirected to exports (see Section 4.3). A recent application on Kuwait using a CGE model capturing its
oil-dependent features matches some of our findings on how energy price reforms counter negative oil price shocks
and foster economic diversification, thus boosting economic growth (Shehabi, 2020).
4.2 Public accounts dynamics
After running large surpluses during the high-oil-prices decade of 2004 to 2013, at almost 12% of GDP on average,
the Saudi public budget deteriorated from 2014 on and recorded a 15.8% deficit in 2015 (SAMA, 2018). This
makes the estimation of budget-related variables a crucial scenario result. The level of sectoral and distributional
detail of IMACLIM-SAU offers significantly finer computations of the distributive effects of diversification
policies, including on public accounts, than the rough estimates of Soummane et al. (2019).
Under our set of assumptions regarding public expenditures and income (see Equations A-19 to A-26), 2014 to
2017 budget balance estimates, although uncontrolled by dynamic calibration, are reasonably close to observed
statistics (Figure 2).
17
Figure 2: Public budget balance under Continuity versus Transformation
Sources: SAMA (2018) for 2013-2017 statistics and IMACLIM-SAU simulations. The 2014 public deficit that escapes IMACLIM-SAU
reflects expenditures growing 15% compared to 2013 mainly because of exceptional military expenditures (+21%) and public investment
(+22%) (SAMA, 2018).
After 2017, the balances of both scenarios improve up to 2025, but following contrasted pathways. Under
Transformation, the balance is close to equilibrium in 2024 and reaches its assumed +1% surplus cap in 2025
(Figure 2). Under Continuity, the balance improves at a slower rate to reach its maximum of -2.8% in 2025, i.e.
remains in deficit. It then marks the downturn of the price of oil projected by the IEA (see Section 3.3), which
forbids public income to increase as fast as public expenditures broadly following general activity, to end on a
5.2% deficit in 2030. Conversely, the steep increases of corporate taxes and energy prices allow the Transformation
budget balance to stabilize at its +1% cap up to 2030. The excess surplus keeps in fact increasing and fuels
additional investment reaching 2.9 GDP points in 2030, bringing investment expenses at 29.0% of GDP compared
to 26.1% under Continuity. By 2030, collected taxes amount to 15.3% of Transformation GDP versus -0.4% of
Continuity GDP. The negative tax collection of Continuity points at the cost of maintaining energy subsidies,
which outweigh total tax proceeds.
IMACLIM-SAU records the accumulation of public balances into public debt taking account of the initial net
public credit and of the net interests that it raises. This allows estimating that, under Transformation, the KSA
government will continue holding a consolidated creditor position up to 2030, stabilized at around 21% of GDP
18
from 2025 on when the budget surplus keeps at its +1% cap. At that later period, our assumption of excess fiscal
surplus acting as a growth catalyzer via extra investment opposes further improvement of the net public credit
position. Notwithstanding, the contrast is strong with the Continuity scenario, whose systematic public deficits
erode the net position. Erosion accelerates after 2025, when public deficits start aggravating. Net financial reserves
are consumed by 2027 and the total net debt reaches 14.9% of GDP in 2030 (Figure 3).
Figure 3: Net public debt under Continuity versus Transformation
Sources: SAMA (2018) for 2013-2017 statistics and IMACLIM-SAU simulations.
The poor public-accounts performance of Continuity is all the more striking as Continuity favors energy-intensive
activities, a major source of public income (see Annex B.1 in the Online Appendix). This is an important drawback
of Continuity because the current large public savings serve as a buffer against oil price volatility. The KSA held
725.5 billion USD reserve assets in 2013, almost equivalent to that yearβs GDP. By the end of 2017, the sharp
decline of the oil price from 2015 on had prompted public intervention that had reduced reserve assets by 31.6%,
to under USD 500 billion, and the gross public debt had climbed to 17.2% of GDP after the Kingdomβs first
issuance of international bonds in decades. Reliance on energy-intensive activities without any broad fiscal reform
to expand the tax base turns out to jeopardize the capacity of the Kingdom to maintain such macroeconomic policy.
19
4.3 Energy consumption and CO2 emissions
IMACLIM-SAU projections of the energy consumptions of productive activities reflect exogenous assumptions
on sectoral energy intensities, which we extract from dedicated runs of the KAPSARC Energy Model (KEM) for
the most important of them (see Annex C.2 in the Online Appendix).
Under Continuity, total Saudi energy consumptions rise 90.3% above 2013 level by 2030 (Table 5). The increase
partly results from our assumption that the maintained regulation of energy prices at low levels does not foster any
energy-efficiency gains in non-energy activities. However, it is more than twice that of real GDP (see Section 4.1)
and also reflects structural change as well as the rapid growth of householdsβ consumptions (see Annex C.2 in the
Online Appendix). Its bulk thus comes from energy-intensive industries and the power sector (Table 5).
By contrast, Transformation brings total Saudi consumption 1.9% below 2013 level in 2030, or 243.1 million tons
of oil-equivalent (Mtoe) below Continuity level. Around half the scenario gap comes from curtailed energy
consumptions of CHM (-117.5 Mtoe) and an additional 40% from ELE sector (-89.6 Mtoe). In CHM, abatement
follows on significant output reductions in favor of MAN and C&S. In ELE, the bulk of energy savings stems
from the phase-out of crude oil as primary energy and its replacement by renewablesβmainly solar PV, notably
from 2023 onβand nuclear projected by KEM. Low-carbon options are not explicitly represented in IMACLIM-
SAU but are captured by higher capital intensity and lower fossil intensity of the ELE sector.8 Under our
assumption common to both scenarios of a moderate increase of refining capacities (see Annex C.4 in the Online
Appendix), the combined sectoral energy savings of Transformation result in an additional 25.9 Mtoe or 29.3%
exports of refined products (+4.6% total energy exports) compared to Continuity in 2030.
8 We use reference technology costs of IEA (2016) to compute the capital costs incurred by the investment chronogram projected by KEM.
20
Table 5: Variations of domestic energy consumptions
Sector Continuity
2030 versus 2013
Transformation
2030 versus 2013
2030 variation
from Cont. to Transf.
OIL n/a n/a n/a
GAS n/a n/a n/a
REF +58.5% +54.2% -2.7%
ELE +74.6% -41.8% -66.7%
AGR +25.7% +3.9% -17.4%
MIN +108.6% -15.7% -59.6%
CHM +246.1% -16.4% -75.9%
NMM +43.2% -9.2% -36.6%
MAN +28.7% -12.2% -31.8%
C&S +46.3% +38.6% -5.3%
WTP +40.2% +20.0% -14.4%
ATP +57.3% +22.4% -22.2%
OTP +52.0% +18.3% -22.2%
Households +64.6% +28.2% -22.2%
Total +90.3% -1.9% -48.5%
Source: IMACLIM-SAU simulations. Sectoral non-commercial auto-consumptions are not recorded in the input-output framework of the
model (i.e., oil and gas consumption within the oil and gas industry). From an energy balance perspective, the consumptions of the REF
and ELE sectors are primary consumptions for transformation purposes, while other consumptions are final consumptions.
The energy mix of projected consumptions of the Continuity scenario is almost as carbon-intensive as that of 2013.
Consequently, the 90.3% increase of consumptions over our projection period results in an 82.0% increase in CO2
emissions, which reach 912.8 million tons (Mt) in 2030.9 Comparatively, Transformation significantly abates CO2
emissions, which end 19.3% below 2013 level in 2030 (Figure 4). In line with energy consumptions, the large
increase in Continuity is mainly attributable to ELE and CHM, which account for 40.9% and 23.8%, respectively
(at 373.3 Mt and 217.3 Mt) of total Saudi CO2 emissions in 2030. At that horizon, emissions of these sectors are
76.3% and 234.3% higher than their respective 2013 levels. Conversely, the reduced CO2 emissions of
Transformation largely reflect lower CHM activity and changes of the energy mix of power supply, which cut
down CHM and ELE emissions by 75.2% and 70.9% in 2030 compared with Continuity. In total, Transformation
emissions end at 404.9 Mt, 507.9 Mt (-55.6%) below Continuity levels. The gap between the two scenarios is
9 We track CO2 emissions as products of energy inflows and IPCC (2006) coefficients of kg CO2 per TJ of consumption: 73,300 for crude oil;
56,100 for natural gas and for refined products; 77,400 in NMM and WTP (mainly fuel oil consumptions); 69,300 in CPS and OTP (mainly
gasoline consumptions); 71,500 in ATP (mainly kerosene consumptions) and 73,300 in other sectors. Exceptions are the crude oil inflows into
refineries (no emissions, computed for final uses of refined products) and the petrochemical sector, for which we consider IEA (2019) estimates
of shares of stored carbon: 80% for refined products and 33% for natural gas. With this method, we estimate 2013 Saudi CO2 emissions at
501.6 Mt i.e. 4.8% above the 478.6 Mt reported in the EDGAR database (https://edgar.jrc.ec.europa.eu/overview.php?v=CO2ts1990-2013).
21
almost four times the Saudi NDC target of curbing down CO2 emissions by 130 Mt in 2030 (Saudi Arabia, 2015).
Thus, Saudi Arabia could largely increase its international commitments if it implemented the Transformation
scenario while treating Continuity as the baseline from which reductions are assessed. Such a move would send a
strong positive signal to the global community considering the importance of Saudi positions in international
climate negotiations.
Figure 4: Total CO2 emissions under Continuity versus Transformation
Sources: EDGAR database (https://edgar.jrc.ec.europa.eu) for 2013-2017 statistics and IMACLIM-SAU simulations.
5 CONCLUSION AND POLICY IMPLICATION
In this paper, we investigate the impact of diversification strategies of the Kingdom of Saudi Arabia (KSA) in
response to rapid global climate action inducing depressed oil prices. To do so, we extend the work of Soummane
et al. (2019) into a multi-sectoral framework with explicit representation of the secondary distribution of income
and consecutive debt accumulation.
Over its projection horizon to 2030, our numerical modeling associates the diversification strategy of
Transformation toward non-energy-intensive sectors and fiscal reforms with higher growth and lower
unemployment than Continuity of expansion of energy-intensive industries. By structural change and through
reforming energy prices, Transformation succeeds in containing domestic energy consumptions, whereas
22
Continuity inflates them significantly beyond real economic growth. Also, the energy-price increases of
Transformation induce partial decarbonization of energy supply that roughly leads to stabilization of Saudi CO2
emissions from 2023 on, at a level significantly below that of 2013 and largely below that of Continuity emissions.
Lastly, through cuts in energy subsidies and gradual fiscal reforms, Transformation considerably improves public
budget outlooks by restoring surpluses from 2025 on. In this regard, Continuity proves a dead-end by failing to
hedge against the IEA-projected 2025 downturn of the international price of oil associated with its Sustainable
Development Scenarioβs global peak in oil demand, thus inducing substantial public deficits and the consecutive
accumulation of a net public debt ultimately jeopardizing the Saudi growth model.10
However, our Transformation scenario rests on export impulses fostering activity shifts and on generalized reforms
of the energy sector and taxation system, whose enabling conditions require further investigation. The take-off of
non-oil trade demands increasing the productivity and competitiveness of the Saudi private sector, which has been
a hurdle to past diversification plans. Export-oriented services and industries rely on the capacity of the authorities
to channel investment towards high value-added branches, in which the KSA could benefit from comparative
advantages (e.g., geographical location, size of the domestic market, regional leadership). For the manufacturing
industries, the KSA could rely on technology transfer and integration with its regional market. Energy efficiency
gains not only lessen the impacts of energy-pricing reforms on the costs of producers and the budgets of
households, but also free up additional refined-products export capacity. Efficiency programs must focus on waste
reduction and demand-side management to limit necessary investments, thus reaching efficiency objectives at least
cost.
Finally, IMACLIM-SAU surmises perfect labor mobility across sectors, which contradicts the rigidity of the Saudi
labor market resulting from its high level of regulation, notably on foreign workers. Similarly, ripping the benefits
of Transformation calls for an improvement of education and training programs to meet its skill requirements,
especially for Saudi nationals. Indeed, Vision 2030 highlights the necessity of aligning the educational system with
future market needs and Saudi authorities launched the National Labor Gateway (TAQAT), which aims at
determining the skills and knowledge required by each socio-economic sector (Saudi Arabia, 2016b).
10 The unfolding COVID-19 crisis could accelerate this 2025 tipping point and the compared desirability of the Transformation option. The
crisis is too recent to be reflected in our analyses: at the time of writing, there is strong uncertainty on its immediate consequences and
repercussions in further years, including on oil markets.
23
DECLARATION OF COMPETING INTERESTS
The authors declare that they have no known competing financial interests or personal relationships that could
have appeared to influence the work reported in this paper.
ACKNOWLEDGMENTS
We are grateful to anonymous referees for their comments and suggestions that have helped to considerably
improve the paper; nonetheless, we are, of course, responsible for all errors and omissions. Development and
implementation of IMACLIM-SAU was part of the Ph.D. research of Salaheddine Soummane funded by the
French EDF R&D company under the ANRT convention CIFRE2015/1313. We thank the King Abdullah
Petroleum Studies and Research Center (KAPSARC) in the persons of Walid Matar and Hossa Almutairi for kindly
providing results of KEM simulations, complementary data for this study, and comments on an earlier draft. The
continuous development of the IMACLIM model benefits from the support of the Chair Long-Term Modeling for
Sustainable Development (Ponts Paristech-Mines Paristech) funded by Ademe, GRT-Gaz, Schneider Electric,
EDF, RTE, Total, and the French Ministry of Environment.
REFERENCES
ABB (2015). Saudi Arabia: Energy efficiency report. Available at
https://library.e.abb.com/public/3286e88f2467b11dc1257be80055df20/Saudi%20Arabia.pdf.
Al Kaabi, F. (2016). βThe Nexus between Remittance Outflows and GCC Growth and Inflationβ. Journal of International Business and
Economics. June 2016, Vol. 4, No. 1, pp. 76-85. DOI: 10.15640/jibe.v4n1a7.
Al Yousif, M., & Al Bakr, A. (2017). Tourism Development Effect On Saudi Economic Diversification. SAMA Working Paper, March 2017.
Available at http://www.sama.gov.sa/en-
US/EconomicResearch/WorkingPapers/Tourism%20Development%20Effect%20On%20Saudi%20Economic%20Diversification.pdf.
Alabbadi, N. M. (2012). Energy efficiency in KSA: Necessity and expectations. Communication to The Second Annual Industrialist Workshop,
Yanbu, Saudi Arabia. Available at https://www.slidedoc.us/2nd-annual-indus-workshop-yanbu-seec-may-2012-pptx.
Albassam, B. A. (2015). βEconomic diversification in Saudi Arabia: Myth or reality?β Resources Policy 44, 112β117.
https://doi.org/10.1016/j.resourpol.2015.02.005.
Al-Hawwas, A. (2010). βPolicy experiments for the Saudiβs economy using a Computable General Equilibrium Model (CGE), oil demand and
tariff liberalization effects on the Saudi economyβ. PhD thesis, University of Dundee, Dundee, United Kingdom. Available at:
https://discovery.dundee.ac.uk/files/1692405/Al-Hawwas_phd_2010.pdf
AlKathiri, N., Atalla, T. N., Murphy, F., & Pierru, A. (2020). Optimal policies for managing oil revenue stabilization funds: An illustration
using Saudi Arabia. Resources Policy, 67, 101686. https://doi.org/10.1016/j.resourpol.2020.101686
Al-Thumairi, I. A. (2012). βFixed or flexible exchange rate for Saudi Arabia: optimal solution of CGE modelβ. International Journal of Basic
and Applied Sciences 1 (2), 67-87. DOI 10.14419/ijbas.v1i2.29.
APICORP, Arab Petroleum Investments Corporation (2018). βSaudi energy price reform getting seriousβ. Vol. 03 No. 05, February 2018.
Available at http://www.apicorp-arabia.com/Research/EnergyReseach/2018/APICORP_Energy_Research_V03_N05_2018.pdf
Bauer, N., Mouratiadou, I., Luderer, G., Baumstark, L., Brecha, R. J., Edenhofer, O., & Kriegler, E. (2013). βGlobal Fossil Energy Markets
and Climate Change Mitigation: An Analysis with REMINDβ. Climatic Change, 136(1), 69-82. DOI 10.1007/s10584-013-0901-6.
24
Billmeier, M. A., & Hakura, M. D. (2008). Trade Elasticities in the Middle East and Central Asia: What is the Role of Oil? (No. 8-216).
International Monetary Fund. Available at https://www.imf.org/external/pubs/ft/wp/2008/wp08216.pdf.
Blanchflower, D. G., Oswald, A. J. (2005). βThe wage curve reloadedβ. National Bureau of Economic Research no. w11338. Available at
https://www.nber.org/papers/w11338.pdf.
Blazquez, J., C. Hunt, L., and Manzano, Baltasar (2017). "Oil subsidies and renewable energy in Saudi Arabia: A general equilibrium
approach." The Energy Journal 38. KAPSARC Special Issue. DOI: 10.5547/01956574.38.SI1.jbla.
BrΓ©mond, V., Hache, E., & Mignon, V. (2012). βDoes OPEC still exist as a cartel? An empirical investigationβ. Energy Economics, 34(1),
125-131. DOI 10.1016/j.eneco.2011.03.010.
Cairns, R. D., & Calfucura, E. (2012). βOPEC: Market failure or power failure?β Energy Policy, 50, 570-580. DOI
10.1016/j.enpol.2012.07.058.
Callen, M. T., Cherif, R., Hasanov, F., Hegazy, M. A., & Khandelwal, P. (2014). Economic diversification in the GCC: Past, present, and
future. International Monetary Fund.
CDSI (2014). National accounts 1435/1436 A.H. (2014). Central Department of Statistics and Information. The Kingdom of Saudi Arabia.
Combet, E., Ghersi, F., Lefèvre, J., Le Treut, G. (2014). Construction of hybrid Input-Output tables for E3 CGE model calibration and
consequences on energy policy analysis. GTAP resource #4524, Global Trade Analysis Program, Purdue University, 27 p.
Crassous, R., Hourcade, J.C., Sassi, O., (2006). βEndogenous Structural Change and Climate Targets. Modeling experiments with IMACLIM-
R.β The Energy Journal, Special Issue: 259-276. DOI: 10.5547/ISSN0195-6574-EJ-VolSI2006-NoSI1-13.
De Lauretis, S, Ghersi, and Jean-Michel Cayla, J-M. (2017). "Energy consumption and activity patterns: an analysis extended to total time and
energy use for French households." Applied Energy 206: 634-648. https://doi.org/10.1016/j.apenergy.2017.08.180.
De Santis, R. A. (2003). βCrude oil price fluctuations and Saudi Arabia's behaviorβ. Energy Economics 25 (2), 155-173.
https://doi.org/10.1016/S0140-9883(02)00106-8.
Devaux, P. (2013) βEconomic diversification in the GCC: dynamic drive needs to be confirmedβ BNP Paribas economic research, July-August
2013. Available at http://economic-research.bnpparibas.com/Views/DisplayPublication.aspx?type=document&IdPdf=22570.
Edenhofer, O., Pichs-Madruga, R., Sokona, Y., Farahani, E., Kadner, S., Seyboth, K., ... Kriemann (2014). Climate change 2014: Mitigation
of climate change. Contribution of working group III to the fifth assessment report of the intergovernmental panel on climate change.
Cambridge, United Kingdom and New York, USA: Cambridge University Press.
El Mallakh, R. (2015). Saudi Arabia: Rush to Development (RLE Economy of Middle East): Profile of an energy economy and investment.
Routledge. https://doi.org/10.4324/9781315744414.
Fattouh, B., & El-Katiri, L. (2013). βEnergy subsidies in the Middle East and North Africaβ. Energy Strategy Reviews, 2(1), 108-115.
https://doi.org/10.1016/j.esr.2012.11.004.
Fattouh, B., & Mahadeva, L. (2013). βOPEC: what difference has it made?β Annual Review of Resource Economics, 5(1), 427-443.
https://doi.org/10.1146/annurev-resource-091912-151901.
Ghersi, F. (2015). βHybrid Bottom-up/Top-down Energy and Economy Outlooks: A Review of IMACLIM-S Experimentsβ. Frontiers in
Environmental Science 3 (74), 1-18. https://doi.org/10.3389/fenvs.2015.00074.
Ghersi, F., & Hourcade, J. C. (2006). Macroeconomic consistency issues in E3 modeling: the continued fable of the elephant and the rabbit.
The Energy Journal Special Issue on Hybrid Modeling, 39-61. DOI 10.5547/ISSN0195-6574-EJ-VolSI2006-NoSI2-3.
Gonand, F., Hasanov, F., and Hunt, L. (2019) "Estimating the impact of energy price reform on Saudi Arabian intergenerational welfare using
the MEGIR-SA model." The Energy Journal 40.3. DOI10.5547/01956574.40.3.fgon
Gupta, D., Ghersi, F., Vishwanathan, S. S., & Garg, A. (2019). Achieving sustainable development in India along low carbon pathways:
Macroeconomic assessment. World Development, 123, 104623. https://doi.org/10.1016/j.worlddev.2019.104623.
Gupta, D., Ghersi, F., Vishwanathan, S. S., & Garg, A. (2020). Macroeconomic assessment of Indiaβs development and mitigation pathways.
Climate Policy 20 (7), 779-799. https://doi.org/10.1080/14693062.2019.1648235
Hourcade, J. C., Jaccard, M., Bataille, C., Ghersi, F. (2006). βHybrid Modeling: New Answers to Old Challengesβ. Introduction to the The
Energy Journal Special Issue on Hybrid Modeling, 1-11. Accessible at: https://www.jstor.org/stable/23297043.
Hourcade, J.-C., Ghersi F., Combet, E., Thery D. (2010). Carbon Tax and Equity : The Importance of Policy Design. In Dias Soares, C., Milne,
J., Ashiabor, H., Deketelaere, K., Kreiser, L. (ed.), Critical Issues in Environmental Taxation, Oxford University Press, Oxford, pp. 277-295.
25
HSBC (2018). βNow, next and how for business. Saudi Arabia reportβ. Available at https://www.business.hsbc.com/-
/media/cmb/tfr/saudi-arabia-report-2018.pdf.
Hvidt, M. (2013) "Economic diversification in GCC countries: Past record and future trends."
http://eprints.lse.ac.uk/55252/1/Hvidt%20final%20paper%2020.11.17_v0.2.pdf.
ICAO (2010). βEnvironmental Report 2010β. International Civil Aviation Organization, Montreal. Available at
https://www.icao.int/environmental-protection/Documents/EnvironmentReport-2010/ICAO_EnvReport10-Ch1_en.pdf.
IEA (2009). Transport energy and CO2: Moving toward sustainability. International Energy Agency, Paris, France.
IEA (2015). Energy Balances of non-OECD Countries 2015. International Energy Agency, Paris, France.
IEA (2016). World Energy Outlook 2016. International Energy Agency, Paris, France.
IEA (2017). World Energy Outlook 2017. International Energy Agency, Paris, France.
IEA (2019). CO2 Emissions from Fuel Combustion. International Energy Agency, Paris, France.
IMF (2016a). Economic Diversification in Oil-Exporting Arab Countries. In Annual Meeting of Arab Ministers of Finance. International
Monetary Fund. Available at: https://www.imf.org/external/np/pp/eng/2016/042916.pdf
IMF (2016b). Saudi Arabia: 2016 Article IV Consultation. Press release; staff report; and informational Appendix. Available at:
https://www.imf.org/external/pubs/ft/scr/2016/cr16326.pdf. International Monetary Fund.
IMF (2016c). World Economic Outlook Database, April 2016. International Monetary Fund,
Washington, D.C. Available at https://www.imf.org/external/pubs/ft/weo/2016/01/weodata/index.aspx.
Jadwa (2017). βPetrochemicals and the Vision 2030β. February 2017.
Jadwa (2018). βUpdate: Outlook on Crude Oil Refining in Saudi Arabiaβ. March 2018.
Johansson, D. J., Azar, C., Lindgren, K., & Persson, T. A. (2009). βOPEC strategies and oil rent in a climate conscious worldβ. The Energy
Journal, 23-50. https://www.jstor.org/stable/41323366
Krane. (2017) "Beyond 12.5: The implications of an increase in Saudi crude oil production capacity." Energy Policy 110: 542-547.
https://doi.org/10.1016/j.enpol.2017.08.052.
Krane, J., and Shih Y. H. (2016). βEnergy Subsidy Reform in the Persian Gulf: The End of the Big Oil Giveawayβ. Issue Brief no. 04.28.16.
Rice Universityβs Baker Institute for Public Policy, Houston, Texas.
Le Treut, G., (2017). Methodological proposals for hybrid modelling: consequences for climate policy analysis in an open economy (France).
Available at https://hal.archives-ouvertes.fr/tel-01707559/document.
LefΓ¨vre, J., Wills W., & Hourcade, J-C. (2018) Combining low-carbon economic development and oil exploration in Brazil? An energyβ
economy assessment, Climate Policy, 18:10, 1286-1295. DOI 10.1080/14693062.2018.1431198
Matar W., Echeverri R., Pierru A. (2016) βThe Prospects for Coal-fired Power Generation in Saudi Arabia.β Energy Strategy Reviews 13-14.
https://doi.org/10.1016/j.esr.2016.10.004
Matar, W., Murphy, F., Pierru, A., Rioux, B., & Wogan, D. (2017). Efficient industrial energy use: The first step in transitioning Saudi Arabia's
energy mix. Energy Policy, 105, 80-92. https://doi.org/10.1016/j.enpol.2017.02.029.
MGI (2015). βSaudi Arabia beyond oil: the investment and productivity transformationβ. McKinsey Global Institute, McKinsey & Company,
December 2015. Retrieved from: https://www.mckinsey.com/featured-insights/employment-and-growth/moving-saudi-arabias-economy-
beyond-oil.
Moya, J.A., & Boulamanti, A. (2016). βProduction costs from energy-intensive industries in the EU and third countriesβ. Publications Office
of the European Union: Luxembourg. EUR27729EN. DOI 10.2790/056120.
Okagawa, A., Ban, K. (2008). βEstimation of substitution elasticities for CGE modelsβ. Discussion Papers in Economics and Bus iness 08-16,
Graduate School of Economics and Osaka School of International Public Policy, Osaka, Japan. Available at:
http://www.esri.go.jp/jp/workshop/080225/03_report5_Okagawa.pdf.
Pershing, J., (2000). βFossil fuel implications of climate change mitigation responsesβ. In: Bernstein, L., Pan, J. (Eds.), Sectoral and Economic
Costs and Benefits of GHG Mitigation. Intergovernmental Panel on Climate Change, Bilthoven.
26
Persson, T. A., Azar, C., Johansson, D., & Lindgren, K. (2007). βMajor oil exporters may profit rather than lose, in a carbon-constrained
worldβ. Energy Policy, 35(12), 6346-6353. https://doi.org/10.1016/j.enpol.2007.06.027.
Rutherford, T. F. (1999). βApplied general equilibrium modeling with MPSGE as a GAMS subsystem: An overview of the modeling
framework and syntaxβ. Computational economics, 14(1-2), 1-46. https://doi.org/10.1023/A:1008655831209.
Sassi, O., Crassous, R., Hourcade, J.-C., Gitz, V., Waisman, H., and Guivarch, C. 2010. βImaclim-R: a modelling framework to simulate
sustainable development pathwaysβ. International Journal of Global Environmental Issues, 10: 5β24. DOI 10.1504/IJGENVI.2010.030566
Saudi Arabia (2005). βFirst National Communication of the Kingdom of Saudi Arabia to the UNFCCCβ.
Saudi Arabia (2011). βSecond National Communication of the Kingdom of Saudi Arabia to the UNFCCCβ.
Saudi Arabia (2015). βThe Intended Nationally Determined Contribution of the Kingdom of Saudi Arabia under the UNFCCCβ. Available at
https://www4.unfccc.int/sites/ndcstaging/PublishedDocuments/Saudi%20Arabia%20First/KSA-INDCs%20English.pdf.
Saudi Arabia (2016a). βThird National Communication of the Kingdom of Saudi Arabia to the UNFCCCβ.
Saudi Arabia (2016b). βSaudi Vision 2030β. Available at: http://www.vision2030.gov.sa
SAMA (2018). βAnnual Statistics 2018β. Saudi Arabian Monetary Authority. Retrieved from http://www.sama.gov.sa/en-
US/EconomicReports/Pages/YearlyStatistics.aspx.
Schers, J., Ghersi, F., Lecocq, F., & Grazi, F. (2015). Green Growth and its Implications for Public Policy-The Case of South Africa. Available
at https://www.afd.fr/en/green-growth-and-its-implications-public-policy.
Shehabi, M. (2020). βDiversification effects of energy subsidy reform in oil exporters: Illustrations from Kuwaitβ. Energy Policy, 138, 110966.
https://doi.org/10.1016/j.enpol.2019.110966.
Soummane, S., Ghersi, F. (2020). βThe IMACLIM-SAU modelVersion 1.0β, CIRED Working Paper 2020-77, CIRED, Paris. Available at
http://www.centre-cired.fr/wp-content/uploads/2020/11/cired_wp_2020_77_soummane_ghersi.pdf.
Soummane, S., Ghersi, F. (2019). βHybrid Input-Output tables for Saudi Arabia at the year 2013β, Mendeley Data, V2, doi:
10.17632/wr9mprpj5k.2.
Soummane, S., Ghersi, F., & LefΓ¨vre, J. (2019). βMacroeconomic pathways of the Saudi economy: The challenge of global mitigation action
versus the opportunity of national energy reformsβ. Energy Policy, 130, 263-282. https://doi.org/10.1016/j.enpol.2019.03.062.
Taylor, L., Lysy, F.J. (1979). Vanishing income redistributions - Keynesian Clues about model surprises in the short run. Journal of
Development Economics 6, 11β29. https://doi.org/10.1016/0304-3878(79)90033-6.
Tax Foundation (2018). Corporate Tax Rates Around the World, 2018. Retrieved from: https://taxfoundation.org/corporate-tax-rates-around-
world-2018/.
Van de Graaf, T., Verbruggen A. (2015). βThe oil endgame: Strategies of oil exporters in a carbon-constrained worldβ. Environmental Science
& Policy 54, 456-562. https://doi.org/10.1016/j.envsci.2015.08.004.
Waisman H., Rozenberg J., Hourcade J-C. (2013). βMonetary compensations in climate policy through the lens of a general equilibrium
assessment: The case of oil-exporting countriesβ. Energy Policy 63, 951β961. https://doi.org/10.1016/j.enpol.2013.08.055.
Annex A IMACLIM-SAU FORMULARY
IMACLIM-SAU operates in a dynamic recursive framework where yearly economy-wide equilibria are connected
by accumulation of the capital stock, financial debts and chained price indexes. From a mathematical point of
view, each yearly equilibrium results from the solving of a system of simultaneous non-linear equations:
27
π1(π₯1, π₯2 β¦ , π₯π, π¦
1, π¦
2β¦ , π¦
π) = 0
π2(π₯1 , π₯2 β¦ , π₯π, π¦
1, π¦
2β¦ , π¦
π) = 0
. . .
ππ(π₯1, π₯2 β¦ , π₯π, π¦
1, π¦
2β¦ , π¦
π) = 0
With π₯π a set of π variables, π¦π a set of π parameters and π
π a set of π functions, for some of them linear, for some
of them non-linear, in π₯π. The ππ functions embody constraints of either an accounting nature or a behavioral nature.
The accounting constraints impose themselves on the modeler for the sake of consistency. The behavioral
constraints, quite distinctively, convey the modelerβs views on economic causalities and correlations.
The count of equations and variables depends on whether IMACLIM-SAU models regulated energy prices or not.
Regulated versus reformed energy prices affect equations (32), (57) and (58), which shift from (a) to (b) variants.
Because of indexed notation, each equation covers up to 169 constraints (the intermediate prices of 13 products
into 13 productions). The detail is as follows:
Equations (4), (7), (8), (10), (11), (12), (14), (15), (16), (17), (18), (19), (20), (21), (23), (24), (25), (26),
(27), (28), (31), (32) (a or b), (33), (34), (37), (40), (42), (43), (45), (47), (60), and (61) cover one
constraint each: 32 constraints.
Equation (58b) covers 2 constraints.
Equations (9), (50), (58a) and (59) cover 4 constraints each: 16 constraints in the regulated (a) variant,
12 constraints in the reformed (b) variant.
Equation (13) covers 8 constraints.
Equations (1), (2), (3), (29), (30), (48) cover 9 constraints each (for 9 non-energy goods): 54
constraints.
Equations (5), (6), (22), (35), (36), (38), (39), (41), (44), (46), (49), (51), (53), (54), (55) and (56), cover
13 constraints each (one equation per sector): 208 constraints.
Equation (57b) covers 26 constraints.
Equation (57a) covers 52 constraints.
Equation (52) covers 169 constraints (input-output prices).
28
The version considering regulated energy prices (a variants) thus counts 539 constraints, while that considering
reformed prices (b variants) counts 511 constraints. The following table identifies the 539/511 variables (Var.
count of last-but-one column) matching these numbers of constraints with, when differentiated, the count of the
regulated model on the left side and that of the reformed model on the right side of a slash sign. The table also lists
all parameters of the model, which for most of them are calibrated at base-year level on our hybrid dataset, for
some others stem from other external sources.
Table A.1: IMACLIM-SAU notations
Notation Description Var. Par.
πΆπ Final consumption of good π by households. Consumptions of energy goods
are exogenous (see Annex C.2). Consumption of AGR is exogenous as well
(follows population dynamics).
8 5
π·π Net debt of agent π β {π», πΉ, πΊ, π ππ} (households, firms, public
administrations, foreign agents).
4 0
πΊπΉπΆπΉπ Gross fixed capital formation of agent π β {π», πΉ, πΊ} (households, firms, public
administrations).
3 0
πΊπ Final public consumption of good π. 13 0
πΌπ Final consumption of good π in investment. 13 0
πΎπ Total capital stock in sector π. 13 0
πΏπ Total labor demand from sector π 13 0
πΎπΏπ Value-added πΎπΏ intensity of the production of non-energy good π. 9 0
ππ Imports of good π. Imports of REF are exogenous (Annex C.4), imports of
OIL, GAS and ELE are exogenously set to 0.
9 4
ππΏπ΅π Net lending or borrowing of agent π β {π», πΉ, πΊ, π ππ} (households, firms,
public administrations, foreign agents).
4 0
ππ Pensioned population 0 1
ππ Total population 0 1
ππ Unemployed population 1 0
π πΆ Consumption budget of households 1 0
π π Gross disposable income of agent π β {π», πΉ, πΊ} (households, firms, public
administrations).
3 0
ππ Total supply of good π. 13 0
ππ Export of good π. Exports of GAS and ELE are exogenously set to 0 in both
scenarios. Additionally, exports of OIL are exogenous in the Transformation
scenario (Annex C.4).
11/10 2/3
ππ Domestic output of good π. Outputs of OIL and REF are exogenous in the
Continuity scenario. Output of REF alone is exogenous in the Transformation
scenario (Annex C.4).
11/12 2/1
ππΎπΏπ Parameter of substitution of πΎ to πΏ in good π production. 0 13
ππΎπΏπ Parameter of substitution of πΎ to πΏ in good π production. 0 13
ππ Effective interest rate on the net debt of agent π β {π», πΉ, πΊ}. 0 3
ππΆπ Price of good π for households. See Annex C.1 for the specific assumptions
regarding energy prices.
13 0
ππΊπ Public price of good π. 13 0
ππΌπ Investment price of good π. 13 0
29
ππΏπ Cost of labor input in the production of good π. 13 0
πππ Import price of good π. The import prices of non-energy goods are constant
(non-energy imports are the modelβs numΓ©raire). See Annex C.3 for the
specific assumptions regarding the import prices of energy.
4 9
ππΎπΏπ Price of value-added good πΎπΏ in non-energy sector π. 9 0
πππ Average price of good π supply (output and imports). 13 0
πππ Export price of good π. See Annex C.3 for the specific assumptions regarding
the export prices of energy.
13 0
πππ Output price of good π. 13 0
πππ Price of good π for the production of good π. See Annex C.1 for the specific
assumptions regarding intermediate energy prices.
169 0
π πΌ Investment effort as a share of GDP. 2013 to 2017 efforts are indexed on The
World Bank statistics. From 2018 on, the ratio linearly converges to its 2013-
2017 average in 2030.
0 1
π‘πΈππΆπ Net energy tax per unit of household consumption of good π. 0 13
π‘πΈππΊπ Net energy tax per unit of public consumption of good π. 0 13
π‘πΈππΌπ Net energy tax per unit of good π immobilization. 0 13
π‘πΈπππ Net energy tax per good π consumption in good π production. 0 169
π‘πππΆπ Net other excise tax per unit of household consumption of good π. 0 13
π‘πππΊπ Net other excise tax per unit of public consumption of good π. 0 13
π‘πππΌπ Net other excise tax per unit of good π immobilization. 0 13
π‘ππππ Net other excise tax per good π consumption in good π production. 0 169
πΌππ Technical coefficient, good π intensity of good π. 0 169
πΏππ Scaling factor on transport margins of transport-providing sectors. 1 0
πΏππ Scaling factor on good π exports accounting for the growth trend of Saudi
export markets.
0 1
π π Technical coefficient, capital (write-off) intensity of good π. Exogenous for
energy goods.
9 4
ππ Technical coefficient, labor intensity of good π. Exogenous for energy goods. 9 4
ππ Rate of net operating surplus (mark-up) in the production of good π. 0 13
ππΎπΏπ Parameter of substitution of πΎ to πΏ in good π production. 0 13
ππ Average per capita pensions benefitting the retired population. 1 0
ππ Average per capita transfers benefitting households outside unemployment
benefits and pensions.
1 0
ππ Average per capita unemployment benefits. 1 0
ππΎπΏπ Elasticity of substitution of πΎ to πΏ in non-energy good π production. 0 9
ππππ Elasticity of the contribution of imports into total good π supply to the ratio of
output to import prices.
0 9
ππππ Elasticity of the share of exports into total good π uses to the ratio of import
to export prices (does not apply to exports of GAS and ELE, exogenously
equal to zero).
0 11
ππ€π’ Elasticity of the purchasing power of wages to the unemployment rate. 0 1
ππΆπ Corporate tax rate. 0 1
πππΌ Average annual monetary inflation rate between the calibration year and all
projected years. 0 1
ππΌπ Income tax rate on householdsβ gross disposable income. 0 1
ππΏππ Social contribution (labor tax) rate applicable to wages in sector π. 0 13
30
ππ Saving rate of households. 1 0
ππππΆπ Specific margin on householdsβ consumption of good π. In the Continuity
scenario, the four margins on energy sales adjust to warrant administered
energy prices. In the Transformation scenario, the margins on GAS and OIL
adjust to align domestic prices on international prices while the margins on
ELE and REF are constant parameters (prices are liberalized).
4/2 9/11
πππππ Specific margin on good π exports. Margins on energy exports adjust to
accommodate exogenous export prices of energy goods.
4 9
πππππ Specific margin on good π consumption in good π production. In the
Continuity scenario, the margins on sales of all energy goods to all sectors
adjust to warrant administered prices. In the Transformation scenario, the margins on GAS and OIL adjust to align domestic prices on international
prices while the margins on ELE and REF are constant parameters (prices are
liberalized).
52/
26
117/
143
ππππ Transport margin on the sales of good π. 4 9
ππΆππ Trade margin on the sales of good π. 1 12
πππ΄ππ Value-added tax rate applying to the consumption of good π. 0 13
πππ Output tax rate on the production of good π. 0 13
ππΎπΊπ Share of capital income of sector i accruing to public administrations. 0 13
ππΎπ» Share of total capital income accruing to households. 0 1
ππππ Ratio to GDP of not-elsewhere accounted for transfers accruing to agent π β{π», πΉ, πΊ} (households, firms, public administrations).
0 3
π½πΌ Scaling factor of immobilizations from calibration year. 1 0
π½πΊ Scaling factor of public consumptions from calibration year. 1 0
ππΏ Scaling factor of labor productivity (technical progress) from calibration year. 0 1
πΊπ΅ Adjustment factor inversely affecting imports and exports of the non-energy
good (see Annex B.2). 0 1
πΊπΏ Adjustment factor affecting labor productivity (see Annex B.2). 0 1
πΊπΎ Adjustment factor affecting capital productivity (see Annex B.2). 0 1
πΊπ€ Adjustment factor affecting real wage correlated to unemployment via the
wage curve (see Annex B.2). 0 1
π΅ Trade balance at current prices. 1 0
πΆππΌ Consumer price index evolution from calibration year. 1 0
πππΌ Import price index evolution from calibration year. 1 0
πΊπ·π Gross domestic product. 1 0
πΊπππ Gross operating surplus of sector π. 13 0
πΏ Total active population (labor endowment) in full-time equivalents. 0 1
πππ Specific margin in sector π. 13 0
π Total taxes and social contributions. 1 0
π’ Unemployment rate. 1 0
ππΎ Rental price of capital 1 0
π€ Average net wage across all sectors. 1 0
π€π Net wage in sector π. 13 0
31
A.1 Firms
Producersβ trade-offs
Trade-offs in the production of energy goods πΈ = {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ} are exogenous assumptions based on
KEM and IEA data (see Annex C).
Non-energy productions follow a standard nested production tree. At the bottom of the tree, capital and labor trade
off with a constant ππΎπΏπ elasticity of substitution to form the value-added aggregate πΎπΏπ. The mobilized quantity
of labor πΏπ is however augmented by a productivity factor π, while both the labor and capital inputs are also
adjusted by dynamic calibration factors πΊ (see Annex B.2). Therefore πΎπΏπ = (πΌπΎπΏπ(πΊπΎπΎπ)ππΎπΏπ +
π½πΎπΏπ
(πΊπΏππΏπ) ππΎπΏπ )
1
ππΎπΏπ , with here and elsewhere, for convenience, ππ
= ππβ1
ππ
. Facing prices ππΎ
and ππΏπ
, cost
minimization induces:
β π β πΈ πΏπ = 1
πΊπΏπ(
πΊπΏππ½πΎπΏπ
ππΏπ
)ππΎπΏπ
(πΌπΎπΏπ
ππΎπΏπ (ππΎ
πΊπΎ)
1βππΎπΏπ+ π½πΎπΏπ
ππΎπΏπ (ππΏπ
πΊπΏπ)
1βππΎπΏπ)
β1
ππΎπΏπ πΎπΏπ (A-1)
β π β πΈ πΎπ = 1
πΊπΎ(
πΊπΎπΌπΎπΏπ
πK)
ππΎπΏπ
(πΌπΎπΏπ
ππΎπΏπ (πK
πΊK)
1βππΎπΏπ+ π½πΎπΏπ
ππΎπΏπ (ππΏπ
πΊLπ)
1βππΎπΏπ)
β1
ππΎπΏπ πΎπΏπ (A-2)
All secondary factor intensities are exogenous, taken from either KEM (energy intensities) or constant at
calibration-year value (non-energy intensities). The value-added intensity of non-energy productions is constant
(Leontief assumption):
β π β πΈ πΎπΏπ
ππ=
πΎπΏπ0
ππ0 (A-3)
The absence of proper estimates for Saudi substitution elasticities led to borrow these parameters from the literature
(Okagawa and Ban, 2008).
32
Table A.2: Elasticities of substitution of capital and labor
Sector ππ²π³
OIL 0.139
GAS 0.139
REF 0.046
ELE 0.46
AGR 0.023
MIN 0.139
CHM 0.33
NMM 0.358
MAN 0.046
WTP 0.31
ATP 0.31
OTP 0.31
C&S 0.31
Net lending and borrowing and net financial debt
The firmsβ gross disposable income π πΉ consists of the remainder of the Gross Operating Surpluses (πΊππ) of
sectors, taking account of the shares accruing to households and public administrations, and a share ππππΉ of GDP
as residual transfers, minus interest payments on their net financial debt π·πΉ, at rate ππΉ, and corporate taxes at rate
ππΆπ on their net operating surplus β πππππππ :
π πΉ = β πΊππππ β β ππΎπΊππΊππππ β ππΎπ» β πΊππππ
+ ππππΉ πΊπ·π β ππΉπ·πΉ β ππΆπ β πππππππ (A-4)
The share ππππΉ, the interest rate ππΉ and the corporate tax rate ππΆπ are constant over time at their 2013 calibration
values.
The πΊππ of sector i is the sum of the consumption of fixed capital ππΎ
πΎπ, the net operating surplus ππ ππ ππ and the
specific margins πππ (which do not sum to 0 after the calibration year):
πΊπππ = ππΎ
πΎπ + ππ πππ ππ + πππ (A-5)
33
The sum of specific margins on sector π sales is:
πππ = β πππππ πππ πΌππ πππ + ππππΆπ
πππ
πΆπ + πππππ π
ππ ππ (A-6)
The margins on non-hybrid sales (the sales of those goods without satellite accounts on physical flows, in the case
of IMACLIM-SAU all non-energy goods) are equal to zero. Additionally, for each hybrid good, the sum of margins
on all sales is equal to zero at calibration year, by construction of the IOT.
At projection years, all positive trade and transport margins remain at their calibration values while the negative
margins, which correspond to those sectors providing the underlying trade and transport services (in the case of
IMACLIM-SAU the C&S sector for trade and the C&S, WTP, ATP and OTP sectors for transport), adjust to
warrant accounting balances:
β ππΆππ πππ(β πΌπππ ππ + πΆπ + πΊπ + πΌπ + ππ)π = 0 (A-7)
β ππππ πππ(β πΌπππ ππ + πΆπ + πΊπ + πΌπ + ππ)π = 0 (A-8)
βπ β {πΆππ, πππ, π΄ππ, πππ} ππππ = (1 + πΏππ) ππππ0 (A-9)
The firmsβ investment effort πΊπΉπΆπΉπΉ is equal to total investment net of the investment of households and public
administrations:
πΊπΉπΆπΉπΉ = β ππΌππ πΌπ β πΊπΉπΆπΉπΊ β πΊπΉπΆπΉπ» (A-10)
The net lending or borrowing (NLB) of firms ππΏπ΅πΉ is the difference between the firmsβ disposable income and
investments:
ππΏπ΅πΉ = π πΉ β πΊπΉπΆπΉπΉ (A-11)
The firmsβ net financial debt π·πΉ evolves according to the accumulated NLBsβthe net financial debts of domestic
agents are the only dynamic variables other than the capital stock and the chained price indexes. Monetary inflation
at annual rate πππΌ degrades the real value of the debt. At date t:
π·πΉ,π‘ = (1 β πππΌ) π·πΉ,π‘β1 β ππΏπ΅πΉ,π‘ (A-12)
34
A.2 Households
Consumer trade-offs
Householdsβ final consumption πΆπ are exogenous for energy goods as well as for agricultural goods AGR. For lack
of analysis in the available literature, the remainder of the consumption budget allocates according to the Cobb-
Douglas assumption of constant budget shares:
βπ β π΄ = {ππΌπ, πΆπ»π, πππ, ππ΄π, πΆππ, πππ, π΄ππ, πππ}
ππΆππΆπ
π πΆββ ππΆππΆππβπ΄
=ππΆπ0πΆπ0
π πΆ0ββ ππΆπ0πΆπ0πβπ΄
(A-13)
Income, savings, investment, NLB and net debt
The after-tax gross disposable income of households π π» proceeds from primary factor income, social transfers,
property income and an aggregate of other secondary transfers.
π π» = β π€π ππ πππ + ππΎπ» β πΊππππ + β ππ πππ=π,π,π
+ ππππ» πΊπ·π β ππ» π·π» β ππΌπ π π» (A-14)
Primary factor income comprises the sum of net wages from all economic sectors β π€π ππ πππ and an ππΎπ» share of
gross operating surpluses πΊπππ, directly accruing to households in the form of, mainly, housing rents (imputed or
real). Social transfers involve pensions ππ
ππ, unemployment transfers ππ
ππ and other social transfers ππ
ππ. ππ
stands for per capita transfers and ππ for a target population: exogenous pensioned population ππ, endogenous
unemployed population ππ or exogenous total population ππ. Other transfers form a constant ππππ» share of πΊπ·π
calibrated at base year. They include international remittances, which reach 4.7% of GDP in the case of Saudi
Arabia at our 2013 calibration year (Al Kaabi, 2016). Property income is the interest payment on the net debt π·π»
at rate ππ» resulting from the balance of income from financial assets and interest payments on liabilities. Income
taxes are paid at rate ππΌπ on disposable income π π».
35
Following on our choice of a Johansen closure (see Section 2), Householdsβ savings at rate ππ adjust to balance
investments and savings. The consumption budget of households is equal to the disposable income net of savings:
π πΆ = (1 β ππ) π π» (A-15)
The investment effort of households πΊπΉπΆπΉπ» is indexed on both disposable income π π» and the aggregate
investment effort π πΌ:
πΊπΉπΆπΉπ» = π΄πΊπΉπΆπΉπ» π π» π πΌ (A-16)
With π΄πΊπΉπΆπΉπ» a constant calibrated on base-year values. The net lending or borrowing of households ππΏπ΅π» is the
difference between their disposable income and their consumption and investment:
ππΏπ΅π» = π π» β π πΆ β πΊπΉπΆπΉπ» (A-17)
Similar to firms, the net household debt at date π‘ resulting from the accumulation of NLBs is:
π·π»,π‘ = (1 β πππΌ) π·π»,π‘β1 β ππΏπ΅π»,π‘ (A-18)
A.3 Public administrations
Public income
The gross disposable income of public administrations π πΊ derives from taxes and social security contributions π,
exogenous ππΎπΊπ and ππππΊ shares of the GOS of sectors (reflecting public participations) and GDP, corrected from
transfers to households β ππ πππ and interest payments at rate ππΊ on the net public debt π·πΊ:
π πΊ = π + β ππΎπΊπ πΊππππ + ππππΊ πΊπ·π β β ππ πππ=π,π,π β ππΊ π·πΊ (A-19)
Tax revenue π comprises primary factor and output taxes, value-added and excise taxes, the income tax and other
direct taxes, and the corporate tax:
π = β ππΏππ π€π ππ ππ + πππ πππ ππ +πππ΄ππ
1+πππ΄ππ
(ππΆπ πΆπ + ππΊπ
πΊπ + ππΌπ πΌπ)
ππ=1
36
+ β β (π‘πΈπππ + π‘ππππ) πΌππ ππππ + (π‘πΈππΆπ + π‘πππΆπ) πΆπ + (π‘πΈππΊπ + π‘πππΊπ) πΊπ + (π‘πΈππΌπ + π‘πππΌπ) πΌπ
+ππΌπ π π» + π‘π» πΆππΌ ππ + ππΆπ β ππ πππ πππ (A-20)
Public expenditures and budget balance
The value of total public consumption is a constant π πΊ ratio to πΊπ·π:
β ππΊπ
πΊππ = π πΊ πΊπ·π (A-21)
Sectoral public expenses grow homothetically from calibration year:
πΊπ = π½πΊ
π΄πΊπ (A-22)
With π΄πΊπ a set of constants calibrated at base year.
Social transfers per capita, ππ
, ππ and π
π, are indexed on the average wage:
ππ
= π΄ππ π€ (A-23)
ππ
= π΄ππ π€ (A-24)
ππ
= π΄ππ π€ (A-25)
With π΄ππ, π΄ππ and π΄ππ three constants calibrated at base year.
Public investment is indexed on total investment. This effectively maintains the public contribution to investment
at calibration-year level (37.7%), which is close to the observed average between 2010 and 2016 (SAMA, 2018)
at 36.2% of total investment.
πΊπΉπΆπΉπΊ = = π΄πΊπΉπΆπΉπΊ β ππΌπ
πΌππ (A-26)
With π΄πΊπΉπΆπΉπΊ a constant calibrated at base year. Similar to firms or households, the NLB of public administrations
is the difference between disposable income and investment:
37
ππΏπ΅πΊ = π πΊ β πΊπΉπΆπΉπΊ (A-27)
The public debt accumulates as:
π·πΊ,π‘ = (1 β πππΌ) π·πΊ,π‘β1 β ππΏπ΅πΊ,π‘ (A-28)
A.4 International trade and the foreign agent
For all energy goods, imports are exogenous, and exports are either exogenous as well, or flow from market
clearing (see below). For the non-energy goods, the share of imports ππ in total resource ππ has a ππππ elasticity
to terms-of-trade and is corrected by the inverse of the export dynamic calibration factor πΊπ΅ (see Annex B.2):
β π β {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ} ππ
ππ=
1
πΊπ΅π΄ππ
(πππ
πππ
)πππ
(A-29)
with π΄ππ constants calibrated on 2013 data. We follow IMF (2016b) using elasticities from Hakura and Billmeier
(2008) to set ππ at -0.09 for all non-energy sectors indistinctly. We regard this elasticity as compatible with the
import structure of the Kingdom, composed of goods with very few domestic substitution opportunities.
Non-energy exports ππ are elastic to terms of trade around exogenous trends πΏππ reflecting the growths of Saudi
export markets as well as diversification strategies (see Section 3). Like import elasticities, we derive πππ from
IMF (2016b) based on Hakura and Billmeier (2008) estimating the elasticity of non-oil exports at 0.69:
β π β {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ} ππ = πΊπ΅ (1 + πΏππ) π΄ππ(
πππ
πππ
)πππ
(A-30)
They are adjusted by πΊπ΅ following dynamic calibration from 2014 to 2017 (see Section 2.2). π΄ππ are another set
of constants calibrated in 2013. The trade balance π΅ is:
π΅ = β πππ ππ β πππ
πππ (A-31)
Both the long-lasting peg of the Saudi riyal to the US dollar and the sensitivity of Saudi exports to the world oil
price forbid considering that real effective exchange rate (REER) variations balance Saudi trade (see Section 2).
Our Continuity scenario rather constrains the REER to reflect the significant statistical relationship between the
38
REER and the trade balance contribution to GDP detected and explained by Soummane et al. (2019). To specify
the relationship, we tested several functional forms including a linear link (with an R2 of 0.622), with little impact
on model results. We settle on an exponential form, which exhibits an R2 of 0.674. This relationship defines the
REER as an exponential function of the trade-balance-to-GDP ratio:
πΆππΌ
πππΌ= π΄π πΈπΈπ + π΅π πΈπΈπ π
πΆπ πΈπΈπ π΅
πΊπ·π (A-32a)
with π΅π πΈπΈπ and πΆπ πΈπΈπ calibrated on 1986 to 2015 statistical observation of the two variables (see Figure 1 of
Soummane et al., 2019), and π΄π πΈπΈπ the adjustment that allows fitting 2013 data.
In our Transformation scenario we drop the constant REER assumption to rather acknowledge the impact of the
massive increase of regulated energy prices on the REER by constraining the price of value-added (πΎπΏ) in the
C&S sector on the same πΏππΎπΏ trajectory that it follows in our Continuity scenario (implicitly, relative to the
numΓ©raire of IMACLIM-SAU i.e. the basket of non-energy foreign goods):
ππΎπΏ_πΆππ
= (1 + πΏππΎπΏ) π·π πΈπΈπ , (A-32b)
with π·REER the value of ππΎπΏ
at calibration year.
The Rest of the world (ROW) agent balances out trade (by selling imports β ππππ ππ and buying exports β π
πππ ππ),
property income and interest payments. Its net lending or borrowing capacity ππΏπ΅π ππ is thus:
ππΏπ΅π ππ = β ππππ ππ β β π
πππ ππ β β ππ π·ππ=π»,πΉ,πΊ β β ππππ πΊπ·ππ=π»,πΉ,πΊ (A-33)
The net debt of foreign agents π·π ππ balances out domestic assets and liabilities:
π·π ππ = β β π·ππ=π»,πΉ,πΊ (A-34)
39
A.5 Market clearings
Goods markets
The balance of goods markets is between resources, which comprise domestic production ππ and imports ππ, and
uses, which consist of the consumptions of all sectors β πΌππ πππ , householdsβ and public consumptions πΆπ and πΊπ,
immobilizations πΌπ and exports ππ. For energy goods, the data hybridization process results in this equation being
expressed in thousand tons-of-oil-equivalent (ktoe), in consistency with the 2013 Saudi energy balance of the IEA.
The public consumptions and immobilizations of all energy goods are equal to zero at calibration year and remain
so up to projection horizons by national accounting convention for the former and by definition for the latter.
ππ = β πΌπππ ππ + πΆπ + πΊπ + πΌπ + ππ (A-35)
ππ = ππ β ππ (A-36)
Labor market
On the labor market, a βwage curveβ describes the elasticity of real wage (the purchasing power of wage π€) to
unemployment π’. The real wage π€/πΆππΌ attached to unemployment at 2013 level (5.6%) is defined as the 2013
average real wage multiplied by labor productivity increase π and a wage moderation factor πΊπ€ (see Annex B.2)
via the calibration of one constant π΄π’:
π€
πΆππΌ= π πΊπ€ π΄π’ π’ππ€π’ (A-37)
The net wages in all sectors evolve in parallel to π€:
π€π = π΄π€π π€ (A-38)
The cost of labor is equal to the wage increased by labor tax contributions:
ππΏπ
= (1 + ππΏππ) π€π (A-39)
Labor demands of all sectors and unemployment balance out labor endowment πΏ:
40
β πΏππ = (1 β π’) πΏ (A-40)
For each sector, labor consumption and output are conventionally related via labor intensity:
πΏπ = ππ ππ (A-41)
The unemployed population ππ is:
ππ = π’ πΏ (A-42)
Capital markets
On the capital market, sectoral demands balance out capital endowment πΎ:
β πΎππ = πΎ (A-43)
With for each sector, similarly to labor:
πΎπ = π π ππ (A-44)
Investment
Investment expenses β ππΌπ
πΌππ form an exogenous share π πΌ of πΊπ·π (investment in energy goods is equal to zero
except for stock variations that are cancelled out in the data-hybridization process).
β ππΌπ
πΌππ = π πΌ πΊπ·π (A-45)
The sectoral structure of investment remains unchanged from the base year to projected horizons:
πΌπ = π½πΌ π΄πΌπ (A-46)
with π΄πΌπ constants calibrated on 2013 data.
GDP
πΊπ·π is defined on the expenditure side as:
41
πΊπ·π = β ππΆππΆπ + ππΊπ
πΊπ + ππΌππΌπ + πππ
ππ β ππππππ (A-47)
A.6 Producer and consumer prices
For non-energy goods, the price of the value-added aggregate ππΎπΏπ
is a canonical function (πΎπΏπ being a CES
product of πΎπ and πΏπ) of prices ππΎπ
and ππΏπ
and of the elasticity of substitution of the two inputs ππΎπΏπ:
β π β {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ}
ππΎπΏπ
= (πΌπΎπΏπ
ππΎπΏπ (ππΎπ
πΊπΎπ
)1βππΎπΏπ
+ π½πΎπΏπ
ππΎπΏπ (π€π
πΊπΏπππ
)1βππΎπΏπ
)
1
1βππΎπΏπ (A-48)
The output or producer price of goods π πππ
is the sum of input costs, output taxes at a πππ rate, and is subject to a
mark-up rate ππ corresponding to the rent on natural resources and/or the net operating surplus:
πππ
= β πππ
πΌπππ + ππΏπ
ππ + ππΎ
π π + πi πππ+ πππ
πππ
(A-49)
The import prices of non-energy goods are exogenous and constant (these goods act as the collective numΓ©raire of
the model). The import prices of energy goods follow exogenous trajectories that are indexed on the price of value-
added in the C&S sector, to account for the differentiated impact of oil-price variations on the Saudi and foreign
economies (see Annex C.3):
β π β {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ} πππ = π΄πππ ππΎπΏ_πΆππ (A-50)
With π΄πππ the ratios of the (year-specific) exogenous international energy prices and the price of value-added in
the C&S sector at base year.
The price πππ
of the total resource in good π, ππ, is inferred from:
42
πππ
ππ = πππ
ππ + πππ
ππ (A-51)
Turning to purchasersβ prices, the price of good π for the production of good π, πππ
, is equal to the resource price of
good π augmented from commercial margins ππΆππ, transport margins ππππ
, agent-specific margins πππππ, energy
taxes π‘πΈπππ and other excise taxes π‘ππππ:
πππ
= πππ
(1 + ππΆππ+ ππππ
+ πππππ) + π‘πΈπππ + π‘ππππ (A-52)
The consumer prices of households, public administrations, the investment good and exports are constructed
similarly with additional value-added taxes (except exports) but drop the unnecessary specific margins when
energy is not concerned (public consumption, investment):
ππΆπ
= (πππ
(1 + ππΆππ+ ππππ
+ ππππΆπ) + π‘πΈππΆπ + π‘πππΆπ) (1 + πππ΄ππ) (A-53)
ππΊπ
= (πππ
(1 + ππΆππ+ ππππ
) + π‘πΈππΊπ + π‘πππΊπ) (1 + πππ΄ππ) (A-54)
ππΌπ
= (πππ
(1 + ππΆππ+ ππππ
) + π‘πΈππΌπ + π‘πππΌπ) (1 + πππ΄ππ) (A-55)
πππ
= πππ
(1 + ππΆππ+ ππππ
+ πππππ) (A-56)
Additionally, the exogenous prices of some energy goods are indexed on the price of value-added in the C&S
sector (by adjustment of specific margins) following Annex C.1.
In scenarios of continued energy-price regulation, e.g. our Continuity scenario of Section 3.1, all domestic energy
prices are regulated:
β π β {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ} πππ = π΄πππ ππΎπΏ_πΆππ (A-57a)
β π β {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ} ππΆπ = π΄ππΆπ ππΎπΏ_πΆππ (A-58a)
With π΄πππ and π΄ππΆπ parameters computed as the ratios of the (year-specific) regulated energy prices and the price
of value-added in the C&S sector at base year (see Annex C.1).
43
In scenarios of energy pricing reforms like our Transformation scenario of Section 3.2, only the prices of OIL and
GAS are anchored to exogenous assumptions, which shift from the historical low regulated prices to international
references (see Annex C.1):
β π β {ππΌπΏ, πΊπ΄π} πππ = π΄πππ ππΎπΏ_πΆππ (A-57b)
β π β {ππΌπΏ, πΊπ΄π} ππΆπ = π΄ππΆπ ππΎπΏ_πΆππ (A-58b)
With π΄πππ and π΄ππΆπ defined as above. The household prices of OIL and GAS are only set up for the sake of
consistency because both underlying consumptions are currently equal to zero and remain so in all scenarios (see
Annex C.2).
Irrespective of energy pricing reforms, energy export prices are at exogenous values indexed on the price of value-
added in the C&S sector (see Annex C.3):
β π β {ππΌπΏ, πΊπ΄π, π πΈπΉ, πΈπΏπΈ} πππ = π΄πππ ππΎπΏ_πΆππ (A-59)
With π΄πππ the ratios of the (year-specific) exogenous international energy prices and the price of value-added in
the C&S sector at base year.
The consumer and import price indexes πΆππΌ and πππΌ are computed as chained indexes, i.e. from one period to
the next, according to Fisherβs formula:
πΆππΌπ‘ = πΆππΌπ‘β1ββ ππΆπ,π‘πΆπ,π‘β1
β ππΆπ,π‘β1πΆπ,π‘β1
β ππΆπ,π‘πΆπ,π‘
β ππΆπ,π‘β1πΆπ,π‘
(A-60)
πππΌπ‘ = πππΌπ‘β1ββ πππ,π‘ππ,π‘β1
β πππ,π‘β1ππ,π‘β1
β πππ,π‘ππ,π‘
β πππ,π‘β1ππ,π‘
(A-61)
44
Annex B CALIBRATION OF IMACLIM-SAU
B.1 Secondary distribution of income
On top of the hybridization of energy flows,11 we expand the original CDSI and GSTAT supply-and-use table data
by disaggregating total labor costs between labor tax contributions and net labor payments. We base our
disaggregation on Saudi legislation regarding insurance contributions. These comprise the social contributions that
employers pay for their Saudi employees (we derive the share of Saudi employment from SAMA, 2018), which
amount to 10% of the employeeβs salary and are due to the General Organization for Social Insurance (GOSI).
Also, a 2% accident insurance for both national and non-national employees, and a 2% unemployment insurance,
which is shared equally between employers and Saudi employees. We also modify CDSI accounts to represent the
substantial public subsidy on electricity prices to both activity sectors and households. On the side of expenditures,
we split investment among households, public administrations and firms by allocating to households the
βresidential building constructionβ expenses from SAMA (2018); to government, the dedicated series from national
accounts (SAMA, 2018); and to firms, the remainder of total investment from the original input-output table.
The additional data required to specify secondary income distribution among households, firms, public
administrations and foreign agents (the βrest-of-the-worldβ or RoW) are not available from the national accounts
of CDSI (2014). We therefore turn to supplementary sources along the following lines.
We distribute the gross operation surplus (GOS) of sectors across the three domestic agents as follows. Firstly, we
allocate to households the income from the real estate and renting activities sector of the original IOT of CDSI.
Secondly, we assume that public authorities capture:
85% of the GOS from oil and gas extraction activities, corresponding to the upper bound of the prevailing
taxation applied by the Saudi government to this branch;
71% of the GOS of the refining sector, corresponding to the share of the public Aramco company in the
Saudi refining capacity;
81% of the GOS of the electricity sector, corresponding to the governmentβs share in Saudi Electricity
Company;
11 Which extends to energy taxes and subsidies (see Soummane et al., 2019).
45
50% of the GOS of mineral activities, corresponding to the governmentβs share in the national company
Maβaden;
And 70% of the GOS of petrochemical activities, corresponding to the governmentβs share in SABIC.
Firms collect the remainder of the total GOS as indicated by CDSI. The resulting distribution of GOS is of 16.4%
to households, 44.7% to the government and the remaining 38.9% to firms.
Concerning direct taxes, corporate taxes apply at a rate of 20% on profits accruing to shareholders of other
nationalities than those of the Gulf Cooperation Council (GCC). For GCC shareholders (including Saudi ones),
there is a 2.5% zakat on profits. Although we already isolate energy-related activities, it remains challenging to
distinguish activities attributable to non-GCC investors. Consequently, we retain only the 2.5% Zakat rate to
compute corporate tax payments accruing to the government. Turning to households, there is currently no income
tax in force in Saudi Arabia. However, there is a 2.5% Zakat tax, which we apply to householdsβ disposable
income.
Concerning social transfers, we calibrate unemployment transfers from public administrations to households on
governmental aid in the framework of the βHafizβ program from the Human Resource Development Fund. We
assume that the 1.11 million job seekers reported by SAMA (2018) for the year 2013 perceived the monthly
financial aid of SAR 2,000. Similarly, we equate total pension disbursements from public administrations to
households to the sum of pension payments and compensations to civilian and military personnel from SAMA
(2018), which reflects data from the Public Pension Agency. For the remainder of social transfers, we consider
total transfers from central government reported by Oxford Economics, to which we subtract the above explicit
transfers.
Property incomes of the three domestic agents correspond to interest payments (or revenues) on net debt positions
(which evolve with the accumulation of net lending or borrowing positions) and thus require specifying interest
rates π, which we assume at 5% for firms and households. Then, the property income is calculated as follows: (i)
for households and firms it corresponds to the product of the debt level (see below for calculation) and the interest
rate; (ii) for public administration, we use the governmentβs βother revenueβ figures from SAMA for the year 2013,
to which we subtract 81% of the ELE operating surplus, the perceived income tax and other taxes. The computed
public property income yields an apparent interest rate of 1.3%. The property income of the RoW balances out the
sum of domestic property incomes.
46
We calibrate the net public debt position at our 2013 base year as the sum of the reserve assets reported by SAMA
(2018), comprised of investment in foreign securities (71.6%), foreign currency and deposits abroad (26.3%),
Special Drawing Rights (1.3%), and reserve position at the IMF (0.7%); net of the total public debt, amounting to
2.1% of GDP during that year. The householdsβ debt for the year 2013 corresponds to the difference between
outstanding personal loans, net of assets of investment funds, bank deposits and quasi-monetary (assuming a share
of 70% for personal purpose), and bank claims. The ROW debt corresponds to the net international investment
position from the balance of payments (BoP) minus gold reserves (SAMA, 2018). The firmsβ debt balances the
total debt of agents.
Finally, we compute an aggregate of remaining βother transfersβ as follows. For households, we use the series of
βPersonal transfersβ, corresponding to workersβ remittances, from the Saudi BoP (SAMA, 2018). For public
administration, we compute βother transfersβ as the difference between the aggregate budget balance and all
resources and expenditures elsewhere accounted for. For the rest of the world, we sum up the opposite of workersβ
remittances and other net current transfers (i.e., credit minus debit) from the BoP and the governmentsβ secondary
income from the BoP. The firmsβ βother transfersβ simply balance out the βother transfersβ of the other three agents.
B.2 Calibration on 2014 to 2017 macroeconomics
Dynamic calibration of IMACLIM-SAU on years 2014 to 2017 targets the main macroeconomic indicators of
GDP, the unemployment rate and the trade balanceβsee Annex D.2 of Soummane et al. (2019) for the detailed
procedure. Adjustment factors impacting capital productivity (πΊπΎ), labor productivity (πΊπΏ), the equilibrium wage
(πΊπ€) and exports and imports (πΊπ΅) are assumed to converge to their 2014-to-2017 averages by 2030. The resulting
factors remain within 7.5% of their 2013 values for those that concern labor, capital, and real wage expectations.
They reach 26.4% for the non-energy trade factor πΊπ΅, which reflects the fact that non-energy trade, although
dwarfed by oil trade, must compensate any statistical discrepancy between our sources for the oil price and exports
on one side (IEA data), and the aggregate trade balance contribution to GDP on the other side (World Bank data).
47
Table B.1: Adjustment factors resulting from 2014-to-2017 calibration
2014 2015 2016 2017 2020 2025 2030
πΊπΎ 0.964 0.962 1.003 0.975 0.976 0.978 0.981
πΊπΏ 1.011 1.027 1.075 1.065 1.058 1.047 1.036
πΊπ€ 1.005 1.011 1.021 1.023 1.020 1.016 1.012
πΊπ΅ 0.736 0.790 1.012 1.104 1.061 0.992 0.928
Note: Calibrated values appear in bold script, projections to 2030 for selected years in light script.
Annex C ENERGY SCENARIOS
Parameterization of the energy consumptions and costs constraining IMACLIM-SAU projections are a
combination of outputs from the Riyadh-based KAPSARC Energy Model (KEM) and complementary exogenous
sources available for the KSA or the broader Middle East region. The hybrid nature of IMACLIM-SAU calibration
data warrants the consistent combination of assumptions.
C.1 Domestic energy prices
After a long period of stagnation of regulated energy tariffs, Saudi Arabia recently engaged in a wide reform of
energy pricing (APICORP, 2018). In 2016, the first phase of the reform increased natural gas tariffs by 67%, the
price of ethane by 133%, and that of refined products (depending on fuel grade) between 50% and 79%. In 2018,
the second phase of the reform targeted householdsβ consumptions. It further increased the price of gasoline
between 83% and 127% (depending on fuel grade) and that of residential electricity by 260% (for consumptions
lower than 6,000 kWh per month).
Our Continuity scenario builds on the assumption that domestic tariffs remain constant in Saudi Riyals after the
two pricing reforms of 2016 and 2018, up to 2030 (Table C.1). Because international non-energy goods act as the
collective numΓ©raire of IMACLIM-SAU (all their relative prices are fixed across time for all scenarios), and to
take account of the differentiated impact of oil price variations on inflation in Saudi Arabia and abroad, particularly
in the US, we follow Soummane et al. (2019, see Annex B.2) by indexing the forced exogenous tariffs on the price
of value-added in the C&S sector (see Equations 57 and 58).
However, Saudi authorities are planning further reforms (IMF, 2016b; Jadwa, 2018), although they have not
communicated target prices. Indeed, energy-pricing reforms are implemented in other Gulf countries, and have
48
accelerated after the sharp decline of the price of oil in 2015 to contain the consecutive massive budget deficits.
Saudi plans may extend to raising energy prices to international levels, as ongoing reforms in Qatar and United
Arab Emirates suggest (Krane and Shih, 2016). Our Transformation scenario reflects such assumptions. It assumes
that domestic oil and gas prices (e.g., for power generation or water desalination) converge with international
prices by 2030. Domestic oil prices consequently rise from 6.35 USD per barrel in 2017 to 69 USD per barrel in
2030, and domestic natural gas prices from 1.25 USD per million British thermal unit (MBtu) to around 4 USD
MBtu (in constant 2016 dollars). These reforms annihilate opportunity costs (the national oil company becomes
indifferent between selling oil domestically or abroad) and foster efficiency gains.
Table C.1: Assumptions on domestic oil and gas prices
In SAR/ton-of-oil
equivalent
Calibration
2013
------Continuity------ ------Transformation------
2030 AAGR 2030 AAGR
πππΌπΏ_π πΈπΉ 116.5 152.7 +1.6% 2,108.7 +18.6%
πππΌπΏ_πΈπΏπΈ 108.6 131.5 +1.1% 1,832.1 +18.1%
πππΌπΏ_πππ 116.5 152.7 +1.6% 2,108.7 +18.6%
πππΌπΏ_ππ΄π 116.5 152.7 +1.6% 1,965.5 +18.1%
ππΊπ΄π_πΈπΏπΈ 111.6 163.5 +2.3% 630.9 +10.7%
ππΊπ΄π_ππΌπ 111.6 163.4 +2.3% 630.9 +10.7%
ππΊπ΄π_πΆπ»π 111.6 199.5 +3.5% 767.2 +12.0%
ππΊπ΄π_πππ 111.6 163.4 +2.3% 630.9 +10.7%
ππΊπ΄π_ππ΄π 111.6 163.4 +2.3% 630.9 +10.7%
Sources: see Annex C.1. The price of energy π input into production π is ππ_π, with sector codes those of Table 1. Unreported prices point at
non-existing consumptions (e.g., no crude oil consumption by C&S sector or by households). AAGR is the Average Annual Growth Rate.
The exogenous prices of oil and gas affect the supply costs of refined products and electricity via the hybrid input-
output matrix. Under Transformation, on top of increased oil and gas prices, we assume reductions of the negative
margins on the sales of refined products (which reflect the differential between the average resource price and the
consumption prices of each sector and households) reaching 50% in 2030. Likewise, we assume cuts on subsidies
to electricity sales to both firms and households reaching 50% by 2030. Both parameters remain constant
throughout our projection horizon under Continuity.
Additionally, we adjust the capital intensity of the ELE sector to capture the impact of changes of the energy mix
backing power supply. Under continued energy-pricing regulation, KEM projects the power mix to remain based
on fossil fuels, although forecasting a gradual shift toward natural gas. Under reformed energy prices, it projects
the power mix to shift to solar photovoltaic (PV) and nuclear sources (Matar et al., 2016). Using capital expenditure
49
and operation and maintenance costs from Matar et al. (2016) and IEA (2016), and accounting for the low rates of
capacity use at base-year as well as for the extra costs of handling intermittency in the case of solar PV, we translate
KEM projections into a gradual decrease of the capital intensity of ELE in our Continuity scenario, reaching 20%
in 2030 compared with base year; and, conversely, in a gradual increase of the capital intensity of ELE in our
Transformation scenario, reaching 18% in 2030 compared with base year.
Table C.2 reports the refined fuels (REF) and electricity (ELE) prices resulting from the above assumptions.
Table C.2: Projected consumer prices of refined products and electricity
In SAR/ton-of-oil
equivalent
Calibration
2013
------Continuity------ ------Transformation------
2030 AAGR 2030 AAGR
ππ πΈπΉ_πΈπΏπΈ 85.4 147.2 +3.3% 1,183.0 +16.7%
ππ πΈπΉ_πΆπ»π 111.6 174.8 +2.7% 1,205.9 +15.0%
ππ πΈπΉ_πππ 49.5 77.6 +2.7% 1,151.7 +20.3%
ππ πΈπΉ_ππ΄π 65.4 102.4 +2.7% 1,165.5 +18.5%
ππ πΈπΉ_πΆππ 654.1 671.4 +0.2% 1,679.4 +5.7%
ππ πΈπΉ_πππ 49.5 76.0 +2.5% 1,151.7 +20.3%
ππ πΈπΉ_π΄ππ 679.7 742.3 +0.5% 1,701.7 +5.5%
ππ πΈπΉ_πππ 274.4 466.5 +3.2% 1,348.0 +9.8%
ππΈπΏπΈ_π πΈπΉ 1,471.7 1,510.7 +0.2% 3,106.0 +4.5%
ππΈπΏπΈ_π΄πΊπ 1,183.0 1,214.4 +0.2% 2,724.0 +5.0%
ππΈπΏπΈ_ππΌπ 1,471.7 1,812.9 +1.2% 3,106.0 +4.5%
ππΈπΏπΈ_πΆπ»π 1,471.7 1,812.9 +1.2% 3,106.0 +4.5%
ππΈπΏπΈ_πππ 1,471.7 1,812.9 +1.2% 3,106.0 +4.5%
ππΈπΏπΈ_ππ΄π 1,471.7 1,812.9 +1.2% 3,106.0 +4.5%
ππΈπΏπΈ_πΆππ 1,490.7 2,540.2 +3.2% 3,131.2 +4.5%
ππΆ_π πΈπΉ 654.1 1,924.9 +6.6% 1,763.4 +6.0%
ππΆ_πΈπΏπΈ 581.5 1,796.4 +6.9% 4,891.0 +13.3%
Sources: IMACLIM-SAU calibration and simulations. The price of energy π input into production π is ππ_π, with sector codes those of Table
1. Unreported prices point at non-existing consumptions. AAGR is the Average Annual Growth Rate.
C.2 Domestic energy flows
We resort to the KAPSARC Energy Model KEM (see Matar et al., 2017, for an application with energy pricing
reforms in the KSA) to settle the impact of the energy-pricing trajectories of Annex C.1 on the energy
consumptions of six industrial sectors: oil and gas upstream activities, refining, electricity, water, petrochemicals
and cement. These sectors cover 71% of total domestic consumptions in 2013, the calibration year common to
50
KEM and our own IMACLIM-SAU. To address the remainder of Saudi consumptions, we complement KEM
outputs with the following assumptions.
Considering Continuity energy prices, KEM projects an increase of crude oil and natural consumptions into power
generation (includes water desalination in KEM) of 189% and 60% from 2013 to 2030. For the remainder of firmsβ
energy consumptions under Continuity, we assume constant energy intensities (constant amounts of energy input
by a unit of output).
For householdsβ consumptions, we assume that residential electricity demand grows at an average rate of 3.2% up
to 2030, which is 2 percentage points below 2007-to-2016 average and close to the projected 2.8% increase of
electricity demand for the Middle-East region by the IEA (2017) in its NPS. In fact, the first round of tariff reforms
resulted in declines (for the first time) in 2016, 2017 and 2018 compared with historical growth of electricity
consumption, making a structural decline of electricity demand growth plausible compared with its historical trend.
Concerning refined products, we assume that the fleet of light duty vehicles reaches 20 million units by 2030,12
96% of which personal cars reflecting current shares reported by GSTAT. Moreover, we assume that fuel economy
increases to reach average Corporate Average Fuel Economy (CAFE) standards of 17.1 km/l, up from 8 km/l in
2012 (Alabbadi, 2012) and consider IEA estimates of average annual car mileage (IEA, 2009). According to these
assumptions, householdsβ fuel uses increase 60% from 2013 to 2030 or at a 2.8% average annual rate (Table C.3).
Under the energy-pricing reforms of Transformation, KEM projects a phase-out of crude oil uses and a 77%
decline of refined products uses (resulting from a phase-out of diesel uses) for power generation (including water
desalination) between 2013 and 2030. We additionally assume that the hikes of energy tariffs combined with
efficiency measures decrease the high energy intensity of the rest of the economy, reflecting wasteful and excessive
energy consumptions (Fattouh and El-Katiri, 2013).
The multisector nature of IMACLIM-SAU allows differentiating efficiency assumptions by sector and energy
vector based on additional external sources. We assume that the electricity intensity of industrial sectors decreases
by 1.4% annually, corresponding to projected gain by ABB (2015) based on the Saudi National Energy Efficiency
Program. For air and water transport, we consider annual efficiency gains at 2.0% and 1.1% up to 2030. Both gains
12 https://www.onlyelevenpercent.com/energy-efficiency-saudi-arabia/.
51
derive from international benchmarks (see IEA, 2016 and ICAO, 2010). Finally, for industrial processes (i.e.,
consumption of OIL, GAS and REF by industrial branches MIN, CHM, NMM and MAN), we take up Soummane
et al. (2019) aggregate assumption of 3% annual efficiency gains, close the 2.5% annual gains projected by IEA
(2016) in its 450Scenario up to 2040. The weighted average of the above assumptions points at aggregate 2.7%
annual energy-efficiency gains. Soummane et al. (2019) compute a low sensitivity of macroeconomic results to
their 3% assumption by testing alternative 0%, 1% and 2% annual gains (see their Section 5.2).
For householdsβ consumptions, we assume that Transformation pricing reforms allow containing the increase of
residential electricity uses at the level of population growth (+33.2% from 2013 to 2030). Regarding transport, we
maintain the assumptions of the Continuity scenario except for fuel economy, which we assume to reach the upper
bound of CAFE standards of 22 km/l by 2030. As a result, householdsβ fuel uses increase by 24% from 2013 to
2030 compared to 60% under Continuity (Table C.3).
It is important to stress that our assumptions on the energy consumptions of non-energy sectors take the form of
intensities, i.e., consumptions per unit output rather than absolute consumptions. Thus, we take into account any
discrepancy in sectoral activity between IMACLIM-SAU and that of KEM (see Matar et al., 2016 for the activity
assumptions backing KEM).
Table C.3: Assumptions on domestic energy consumptions
Domestic energy use
(index 1 in 2013)
Calibration
2013
------Continuity------ ------Transformation------
2030 AAGR 2030 AAGR
52
πΌππΌπΏ_π πΈπΉ 1.00 1.00 id. 1.00 id.
πΌππΌπΏ_πΈπΏπΈ 1.00 1.71 +3.2% 0.00 -100.0%
πΌππΌπΏ_πππ 1.00 0.86 -0.9% 0.60 -1.9%
πΌππΌπΏ_ππ΄π 1.00 1.00 id. 0.60 -1.9%
πΌπΊπ΄π_πΈπΏπΈ 1.00 0.95 -0.3% 0.93 -0.4%
πΌπΊπ΄π_ππΌπ 1.00 1.00 id. 0.60 -1.9%
πΌπΊπ΄π_πΆπ»π 1.00 1.08 +0.4% 0.60 -1.9%
πΌπΊπ΄π_πππ 1.00 0.96 -0.3% 0.60 -1.9%
πΌπΊπ΄π_ππ΄π 1.00 1.00 id. 0.60 -1.9%
πΌπ πΈπΉ_πΈπΏπΈ 1.00 0.34 -6.1% 0.20 -9.0%
πΌπ πΈπΉ_πΆπ»π 1.00 1.84 +3.6% 0.60 -1.9%
πΌπ πΈπΉ_πππ 1.00 0.90 -0.6% 0.60 -1.9%
πΌπ πΈπΉ_ππ΄π 1.00 1.00 id. 0.60 -1.9%
πΌπ πΈπΉ_πΆππ 1.00 1.00 id. 1.00 id.
πΌπ πΈπΉ_πππ 1.00 1.00 id. 0.83 -1.1%
πΌπ πΈπΉ_π΄ππ 1.00 1.00 id. 0.71 -2.0%
πΌπ πΈπΉ_πππ 1.00 1.00 id. 0.80 -1.3%
πΌπΈπΏπΈ_π πΈπΉ 1.00 1.00 id. 1.00 id.
πΌπΈπΏπΈ_π΄πΊπ 1.00 1.00 id. 0.78 -2.8%
πΌπΈπΏπΈ_ππΌπ 1.00 1.00 id. 0.78 -2.8%
πΌπΈπΏπΈ_πΆπ»π 1.00 1.20 +1.1% 0.50 -2.8%
πΌπΈπΏπΈ_πππ 1.00 1.32 +1.7% 0.78 -2.8%
πΌπΈπΏπΈ_ππ΄π 1.00 1.00 id. 0.78 -2.8%
πΌπΈπΏπΈ_πΆππ 1.00 1.00 id. 0.78 -2.8%
πΆπ πΈπΉ 1.00 1.60 +2.8% 1.24 +1.3%
πΆπΈπΏπΈ 1.00 1.70 +3.2% 1.33 +1.7%
The volume of energy π input into production π is πΌπ_π, with sector codes those of Table 1. Householdsβ consumption of energy good π is πΆπ.
Unreported volumes are non-existent. AAGR is the Average Annual Growth Rate.
C.3 Energy trade prices
Oil trade accounts for 83% of Saudi exports earning, of which 73% are crude oil exports at around 7 million barrels
per day (mb/d) during the past decade (SAMA, 2018). This makes crude oil price the main variable of interest for
energy trade. Although OPEC supplies 40% of world oil demand, with Saudi Arabia acting as leader with 30% of
the Organizationβs supply, its impact on oil price is not established. There is no agreement about OPECβs market
53
power. Many authors argue that the βcartelβ strategy was established only during some periods, and that OPECβs
strategy has been evolving (Fattouh and Mahadeva, 2013). BrΓ©mond et al. (2012) show that OPEC has been acting
as a price taker for most of the period following the first oil shock (1973), and that cartel behavior only concerns
a sub-group of the Organization. Some even argue that the cartel status of OPEC has never existed (Cairns and
Calfucura, 2012). In the light of these claims, we assume that the KSA does not influence world prices and build
both our Continuity and Transformation scenarios on a common exogenous assumption of the world oil price
trajectory. Blazquez et al. (2017) adopt a similar specification.
We take this trajectory from the Sustainable Development Scenario (SDS) of IEA (2017). The dramatic oil-price
decline of the end of 2014 resulted in historically low investment levels (IEA, 2016). Facing increasing global oil
demand peaking only in the mid-2020s, the current under-investment in oil resources lifts the oil price up until
2025. From then on up to 2030, the global penetration of electric mobility and higher efficiency gains in the
transport sector in addition to tightened climate policies cause global oil demand to decline, ending 16.6 mb/d or
16% lower than that of the less ambitious New Policy Scenario (NPS) in 2030. The oil price follows a similar
trend, declining to 69 USD in 2030 or 26.6% below the price of the NPS scenario.
A fraction of Saudi energy exports consists of refined products. We link the price of such products to that of crude
oil via a differential that we assume constant over time and across scenarios. We calibrate the differential as the
ratio of the weighted average of the prices of exported petroleum products to the price of crude oil. The ratio is
around 0.96 in 2013. Its low level stems from the fact that Saudi exports of refined products consist mainly of both
heavy (fuel oil) and extra-light products (LPG and naphtha).13 Similarly to exports, we assume that the average
import price of refined fuels is indexed on oil prices over our projection horizon and across scenarios.
C.4 Energy trade flows and domestic supply
The four energy goods disaggregated by IMACLIM-SAU separate in two groups as regards market balance: one
group has exogenous domestic output and imports, and endogenous exports balancing resources and uses; the other
group has exogenous imports and exports, and endogenous domestic output balancing resources and uses. Under
Continuity, crude oil (OIL) and refined fuels (REF) belong to the former group and natural gas (GAS) and
13 According to the 2013 Saudi energy balance (IEA, 2015), the export mix is 29% of LPG, 26% of naphtha, 24% of fuel oil, 10% of diesel,
9% of kerosene and 2% of gasoline.
54
electricity (ELE) to the latter. Transformation moves OIL from the former to the latter group for reasons explained
below.
Our Continuity scenario builds on the assumption that the Saudi output of crude oil reaches 12.7 mb/d by 2030.
This corresponds to the Saudi oil supply projected by the IEA (2017) in its NPS scenario, i.e. surmises that shifting
from NPS to SDS (whose oil prices sustain both our scenarios, see Annex C.3 above) does not affect Saudi output,
mostly directed to exports, considering its very low costβsee Annex E.4 of Soummane et al., 2019, for further
discussion. Indeed, the KSA is considering reaching this level of output and could already do so by mobilizing its
spare capacity (Krane, 2017). Concerning the output of refined products REF, our two scenarios share the
assumption that the Saudi refining capacity will increase from the current level of 2.9 mb/d to 3.3 mb/d by 2030,
following up on the opening of the Jazan refinery (0.4 mb/d).
Imports of OIL, GAS and ELE are equal to zero in current statistics as well as in the outlooks of the KEM model
sustaining our scenarios. IMACLIM-SAU simulations keep them so at all years and across scenarios. Imports of
REF are not described by KEM and require some exogenous assumption. For lack of sources on the matter, we
assume that they follow potential growth (the growth of efficient labor supply) i.e. increase by 45.7% from 2013
to 2030 under Continuity, while they keep at 2013 levels in the face of strongly abated domestic demand under
Transformation.
Similarly to imports, exports of GAS and ELE are consistently equal to zero in statistics or KEM outlooks (see
Matar et al., 2016). Regarding GAS, although the Kingdom is a major gas producer, all the production is directed
to the domestic market. Regarding ELE, plans for regional market integration start to take effect but the traded
volumes should remain negligible. We therefore keep both exports at zero across years and scenarios.
Additionally to GAS and ELE, the Transformation scenario considers exogenous OIL exports. The reason is that
the substantial crude oil savings induced by energy-pricing reforms would liberate large additional export
capacities under the assumption of maintained output. We rather assume that Saudi Arabia contains crude oil
exports to avoid flooding global oil markets thus further depressing the global oil price (Blazquez et al. 2017), and
adjusts output accordingly. To facilitate scenario comparison, we set the OIL exports of Transformation at the
endogenous levels that they reach under Continuity. This assumption brings crude oil output under Transformation
11.7% below Continuity levels in 2030. The loss of export revenue is partially compensated by endogenous
55
increases of REF exports flowing from maintained refining capacities and abated domestic demand. In 2030,
refined products exports are 29.3% higher under Transformation than under Continuity.
Annex D SENSITIVITY OF TRANSFORMATION TO STRUCTURAL
CHANGE DRIVERS
We test sensitivity of the Transformation scenario to its three major structural change drivers: energy price reforms,
non-energy export boosts and corporate tax take-off.14 We exclude the Continuity scenario from the sensitivity
analysis as this scenario corresponds to the business-as-usual case, which assumes an increase of already
established activities without reforms of energy prices or corporate tax adjustment. Concerning energy prices (E
prices), we run a low variant maintaining them at 2018 levels i.e. Continuity values, and a high variant considering
a 75% rather than a 50% decrease of explicit and implicit energy subsidies at end-horizon (see Annex C.1).
Concerning export trends (X trends) of the two targeted sectors, i.e., MAN and C&S, we run a low variant bringing
down MAN and C&S trends to default +3.4% per year, and a high variant setting the MAN trend at +15% per year
and the C&S trend at +10.3% per year. This compounds into quadrupling C&S exports by 2030 compared with
targeted tripling in Transformation (see section 3.2). Finally, we test corporate tax rates at the lower and upper
bounds of G20 countries, i.e., 19% (low variant) and 35% (high variant). We conduct the tests all other parameters
equal.
Activity (GDP and employment), public accounts and CO2 emissions results appear qualitatively robust to the
tested ranges, with altered Tranformation scenarios systematically more favorable than Continuity (Table D.1).
Activity responds particularly well to increased corporate taxes under the assumption of excess public surplus
financing additional investment. Conversely, public budget balance and the net public debt are stable across
variants because of the assumed 1% cap on budget surplus, which even the low corporate tax variant does not
prevent reaching. CO2 emissions broadly follow activity for export trend and corporate tax variants. They
expectedly respond to energy pricing, although with low (negative) elasticity because of energy intensities kept
constrained at (implicitly controlled) Transformation levels (see Annex C.2).
14 For the sake of concision, we refer to Soummane et al. (2019) for assessment of the influence of the global oil price on Saudi macroeconomic
outlooks with a 2-sector βKLEMβ aggregation of IMACLIM-SAU.
56
Table D.1: Sensitivity of Transformation to structural change drivers at 2030 horizon
Indicator Low E
prices
High E
prices
Low X
trends
High X
trends
Low
corp. tax
High
corp.
tax
Continuity
Real GDP -0.4% +0.3% -0.6% +0.9% -0.5% +1.6% -1.3%
Trade balance, % GDP
-0.1 pts -0.1 pts -2.5 pts +3.2 pts +0.4 pts -1.1 pts +4.3 pts
2013-2030 trade surplus
-2.5% +0.6% -15.4% +16.8% +0.9% -2.9% +19.0%
Unemployment
rate -0.2 pts +0.2 pts +0.1 pts -0.2 pts +0.3 pts -0.9 pts +1.7 pts
Public budget
balance, % GDP
+0.0 pts -0.0 pts -0.0 pts +0.0 pts +0.0 pts -0.0 pts -6.2 pts
Net public debt, % GDP
+1.6 pts -1.1 pts -0.0 pts +0.0 pts +1.1 pts -1.7 pts +35.9 pts
CO2 emissions, Mt
+2.4% -0.7% -0.7% +1.1% -0.4% +1.2% +125.4%
Source: IMACLIM-SAU simulations. βPtsβ stands for percentage points. The table reports deviations from Transformation results in 2030.
The last column places Continuity relative to central parametrization Transformation for comparison purposes.
On trade, the high export trend variant brings the balance and surplus performance of Transformation close to that
of Continuityβwhile it further increases the levels of activity and employment. This confirms the potential
importance of non-oil, non-energy intensive trade in the successful diversification of the Saudi economy. Energy
prices affect trade via contrary effects on the competitiveness of Saudi non-energy products, and the availability
of refined products for exports, considering the adjustments of domestic demand (see Annex C.2). The corporate
tax rate affects trade through the latter effect only, i.e. opposite to domestic refined fuels consumptions, which
follow activity.
Finally, the high energy prices variant increases unemployment by reducing the purchasing power of households.
This points at the necessity to factor in the ability of households to cut down their energy consumptions in reaction
to energy pricing reforms.